Egarg Idaho Under DOE Contract No

r,i If NUFlEGICR-4375 EGG-2415

o National E Operated by the U.S. De

Steven R. Adams

October 1985

ared for the

uclear R atory Commission EGarG Idaho Under DOE Contract No. DE-

6t' ;s ~~~~ DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency Thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

DISCLAIMER N UR EGlCR-4375 This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their EGG-2415 employees, makes any warranty, express or implied, or assumes any legal liability or responsi- Distribution Category: R1, R 7 bility for the accuracy, completeness, or usefulness of any information, apparatus, product, or 7--- process disclosed, or represents that its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views NUREG/CR--Q 3 75 and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. TI86 004392

THEORY, DESIGN, AND OPERATION OF LIQIJID METAL FAST BREEDER REACTORS, INCLUDING OPERATIONAL HEALTH PHYSICS

Steven R. Adams

Published October 1985

Idaho National Engineering Laboratory EG&G Idaho, Inc. Idaho Falls, Idaho 83415

Alan K. Roecklein NRC Project Manager

. Prepared for the Division of Facility Operations Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Under Contract No. DE-AC07-761D01570 FIN No. A6314

~~~~~~~~~j~18 II Oi 'if25 Gu"3fE3ii IS ~~~ ABSTRACT

A comprehensive evaluation was conducted of the radiation protection practices and programs at prototype LMFBRs with long operational experience. Installations evaluated were the Fast Flux Test Facility (FFTF), Richland, Washington; Experimen- tal Breeder Reactor I1 (EBR-11) Idaho Falls, Idaho; Prototype Fast Reactor (PFR) Dounreay, Scotland; Phenix, Marcoule, France; and Kompakte Natriumgekuhlte Kernreak Toranlange (KNK II), Karlsruhe, Federal Republic of Germany. The evaluation included external and internal exposure control, respiratory protection procedures, radiation surveillance practices, radioactive waste management, and engineering controls for confining radiation contamination. The theory, design, and operating experience at LMFBRs is described. Aspects of LMFBR health physics different from the LWR experience in the United States are identified. Suggestions are made for modifications to the NRC Standard Review Plan based on the differences.

FIN No. A6314-Occupational HP at LMFBRs

11 EXECUTIVE SUMMARY

A study funded by the Nuclear Regulatory Com- core is greater in an LMFBR than in ai? LWR. mission (NRC) was conducted to identify and Operational health physics experience at L MFBRs describe the differences between Light Water Reac- with radiation and contamination control has been tor (LWR) and Liquid Metal Fast Breeder highly satisfactory. Personnel exposure are very low (LMFBR) health physics programs. The NRC will in comparison to LWRs. use the information to update the Standard Review Plan. The Nuclear Regulatory Commission’s Gaseous radionuclides present in LMFBR efflu- Standard Review Plan (SRP) is prepared for the ents are Ne-23, Ar-39, Ar-41, noble gases, and trit- guidance of the Office of Nuclear Reactor Regula- ium. The only gaseous radionuclide produced in tion’s staff responsible for review of applications to LMFBRs that exceeds that of an LWR at compara- construct and operate nuclear power plants. The ble thermal power is tritium. Comparison c’f liquid SRP provides the NRC review staff sufficient radioactive waste produced at LMFBRs with that information to evaluate the applications to ensure produced at LWRs shows that the volume and that the nuclear power plant can be constructed activity released at LMFBRs to be orders of magni- and operated without undue risk to the health and tude smaller. However, solid waste disposal at safety of the public. LMFBRs will require more research and development for converting contaminated sodium to inert In a coordinated work effort with Argonne compounds suitable for disposal. National Laboratory-West (ANL-W) in 1983, health physics programs at operating LMFBR facil- Methods used for monitoring liquid and gaseous ities in the United States, Scotland, France, and the effluents at LMFBRs were not found to be different Federal Republic of Germany were evaluated. An than at LWRs. Process monitoring of sodium extensive literature review was also made concern- purity, measuring and recording of radioactive con- ing the theory, design, and operating experience at centrations in plant sodium, and the use of cold past and present LMFBRs. traps for maintaining control of sodium impurities in the liquid coolant system were found to be very The health physics evaluation included defining different from the analogous LWR water chemistry the radiological source terms at LMFBRs and char- technology. The potential role of cold trapping for acterizing them qualitatively and quantitatively removing radioactive products from the coolant of throughout the reactor systems. Investigated were LMFBRs has not been fully explored. A systematic radiation surveillance or monitoring methods, study of cold trapping performance in operating radiation protection practices and programs deal- reactors is highly desirable. ing with external and internal exposure control, waste management, respiratory protection procedures, and engineering controls for confining radi- Sections of the NRC Standard Review Plan ation and contamination. Elements of the health (NUREG-0800) pertaining to occupational health physics programs unique to LMFBRs have been physics will require substantial modification before identified. they can be applied to LMFBRs. The modifications are owing mainly to the use of sodium coolant. The primary differences in the operational health Regulatory Guides 1.8, 1.33, 8.8, 1.112, 8.2, and physics programs at LMFBRs are principally owing 8.10 will require changes in order to be applicable to the use of sodium as a coolant. Use of sodium to LMFBRs. Other acceptance criteria used in the introduces unique radiological source terms, Na-24 Standard Review Plan that will have to be modified and Na-22. Sodium also presents distinct chal- for use in licensing LMFBRs are NURECi-0718, lenges in the handling of radioactive waste and in -0737, -0103, -0123, and -0212. doing maintenance on sodium-wetted equipment. . Personnel at LMFBRs require training in sodium Future research and development needs for technology to meet the special challenges. Since the LMFBRs are related to health physics instrumenta- power density is higher in an LMFBR than in an tion, secondary containment studies, radiological LWR, and the neutrons in an LMFBR are not source term kinetics in containment and the envi- slowed down to thermal energies by the sodium, the ronment, and the use of probabilistic risk assess- potential for high energy neutron leakage from the ment for dose optimization.

... 111 Appendix A discusses in depth the theory, design. The review of operational experience design, and operating experience of LMFBRs. It includes the reactor features and operating history discusses the physics of breeding, major design of early experimental LMFBRs, early power and objectives, and mechanical and thermal systems test reactors, and the prototype reactors.

iv CONTENTS

ABSTRACT ...... ii

EXECUTIVE SUMMARY ...... 111

THE BREEDER REACTOR AND NUCLEAR POWER ...... 1

RADIATION PROTECTION AT LIQUID METAL FAST BREEDER REACTOR PLANTS ...... 2

Operational Health Physics Program Responsibilities ...... 2

ALARAPrograms ...... 2

Shielding ...... 10

Radiological Source Terms ...... 14

Ventilation ...... 18

Design Objectives ...... 18 SourceTerms ...... 19

Radioactive Waste Management ...... 23

Liquid Waste Management ...... 23 n . Gaseous Wastesystems ...... 24 Solid Waste System ...... 29 . Process and Effluent Radiological Monitoring Systems ...... 33 m Sodium Impurity Monitoring and Analysis Systems ...... 39

SodiumCoolant ...... 39 Sodium Cold Trap Technology ...... 40

Training ...... 50

COMPARISON OF LMFBR HEALTH PHYSICS PROGRAM WITH THE REQUIREMENTS OF THE USNRC STANDARD REVIEW PLAN ...... 53

FUTURE RESEARCH REQUIREMENTS ...... 56

REFERENCES ...... 57

APPENDIX A-LIQUID METAL FAST BREEDER REACTOR: THEORY. DESIGN. ANDOPERATIONALEXPERIENCE ...... A-1

FIGURES

1. Tritium and hydrogen concentration in EBR-I1 sodium coolant systems ...... 28

2 . GLASS activities data failure of a metal-fueled driver element at EBR-11 ...... 34

V 3 . Three-dimensional plot of xenon gas tag nodes for the experimental subassemblies in EBR-I1 n during Run 111 ...... 37

4 . Schematic of a sodium cold trap for MASCOT calculations ...... 42 TABLES

1 . Man-rem expenditures per year [MW/y(t)] for LMFBRs and LWRs ...... 5

2 . Man-rem expenditures at the Dounreay Nuclear Power Development Establishment ...... 5

3 . CY-82 personnel exposure summary ...... 6

4 . Summary of routine radiation survey data during FFTF Cycle 1 ...... 8

5 . Exposure histories for EBR-I1 work groups ...... 9 6 . EBR-I1 annual exposure summaries ...... 10

7 . Experimental program related to CRBRP shielding ...... 13

8 . Corrosion product radionuclides ...... 16

9 . FFTF design basis cover gas activity ...... 19

10. ANL-WEST and EBR-I1 radioactive gaseous releases, 1968-1978 ...... 20 4

11 . Comparison of LMFBR liquid radioactive waste production with U.S. PWRs ...... 24

12 . EBR-I1 radioactive airborne effluents ...... 26

13. Summary of EBR-I1 data of tritium distribution ...... 26

14. Methods used to identify sources of fission-product release in EBR-I1 ...... 36

15 . Intercomparison of FBR coolants ...... 41

16. General guidelines for IHTS cold trap design ...... 44

17. Radionuclides detected in primary circuit cold traps of FMFBRs ...... 45

18. Purification coefficients for the BOR-60 cold trap ...... 46

19. Accident conditions categorized by ANS standards ...... 55

vi ...... -.. -

/\ THEORY, DESIGN, AND OPERATION OF LIQUID METAL FAST BREEDER REACTORS, INCLUDING OPERATIONAL HEALTH PHYSICS;

THE BREEDER REACTOR AND NUCLEAR POWER

Early nuclear physicists often speculated on reac- In the development of LMFBRs, it wits recog- tors of the future in terms of their being a source of nized that the potential for accidental release of commercial power. One of their concerns was the radiation and radioactive materials is present as it is long-range availability of U-235 and Pu-239. Fermi in any nuclear reactor system. A considerable por- and Zinn, especially, became intrigued with the tion of the overall technical effort expended within idea of breeding plutonium from U-238, and as the nuclear industry in LMFBR development has early as the spring of 1944 they discussed the possi- been allocated to minimize the potential releases. bility of building a fast-neutron breeder reactor to This safety awareness, which consciously, perme- demonstrate the feasibility of the breeding concept. ates the process from conceptual design of ii reactor Eventually this interest resulted in the development through the licensing process and long-term opera- of the world’s first fast neutron reactor, tion, ensures that LMFBRs will be operated so as to Clementine, at Los Alamos, New Mexico in 1946. minimize any undue risk to the health and safety of By December 20, 195 1, the first usable amounts of facility personnel and the public, and prevent harm electricity to be generated from a nuclear power to the environment. reactor were produced from a liquid metal fast breeder reactor (LMFBR), EBR-1, at Arco, Idaho. In the 34 years after EBR-1, four LMFBRs in the One element of the safety program at LMFBRs is 250- to 600-We power range began producing elec- health physics. Operational Health Physics is con- tricity and desalinizing water in three European coun- cerned with radiological engineering and protec- tries, and construction on a 1200-MWe LMFBR tion, personnel and environmental montoring, neared completion. By the end of this decade it is and plant chemistry. expected that LMFBRs in the 1200- to 1600-MWe range will be under construction in four countries-the United Kingdom, France, West Germany, and the This NUREG reviews the operational health phys- U.S.S.R. Eight nations on three continents will be ics programs at LMFBRs and compares and contrasts operating LMFBR power reactors or fast test reactors them with the programs at light water reactors within the next ten years. For many countries with few (LWRs). The body reviews LMFBR opei-ational indigenous energy resources, the dream that began in health physics program objectives, radiation shield- the middle of the 20th century of a safe, economical, ing, air contamination control, waste programs, inexhaustive, and practically independent energy sodium chemistry, and training. The extensive appen- source appears likely to become a reality early in the dix reviews the theory, design, and operationail experi- 2 1st century. ence at LMFBRs during the last 34 years.

1 RADIATION PROTECTION AT LIQUID METAL FAST BREEDER REACTOR PLANTS

Operational Health Physics The primary responsibilities and objectives of Program Responsibilities the operational health physics programs at an LMFBR facility are no different than those at a Radiation protection management at nuclear light water reactor (LWR) plant. However, the generating stations is responsible for operational organization, personnel, and training required to health physics activities at the plant. Radiation meet these responsibilities are somewhat different management ensures enforcement of radiation at the LMFBR facility. standards and procedures, reviews proposed methods of plant operation, participates in devel- The differences in operational health physics opment of plant documents, and assists in the plant programs at LMFBRs principally result from the training program by providing specialized training use of sodium as the reactor coolant. A few sub- in radiation protection. During preoperational tests stantive modifications in the organization and and after plant start-up, it provides health physics administration of a health physics program coverage for all operations, including maintenance, designed for a LWR are required for its use at an fuel handling, waste disposal, and decontamina- LMFBR. An LMFBR health physics program tion. It is responsible for personnel and in-plant necessitates the factoring of industrial hygiene and radiation monitoring, environmental surveys and safety in the use of sodium into radiation protec- monitoring, and health physics instrument calibration, modifying training requirements to include tion, and maintains records of personal exposures, sodium technology, reviewing the analytical models and in-plant radiation and contamination levels. It used in preoperational siting studies to determine is responsible for ensuring that plant personnel the quantitative effects sodium has on releases to exposures are maintained as low as is reasonably the environment, and modifying, if necessary, the 4 achievable (ALARA) through implementation of environmental monitoring program. an established ALARA program. Personnel education, experience, and training Operational Health physics management is requirements for an LMFBR operational health responsible for providing an adequate program of physics program differ only slightly from those at health surveillance for all plant operations involv- an LWR. LMFBRs require personnel trained in ing potential radiation hazards. They inform the handling and decontaminating sodium wetted sur- plant manager of all radiation hazards and condi- faces, sodium chemistry, and monitoring sodium tions relating to potential exposures, contamina- purity. tion of plant and equipment, or contamination of site and environs. Their responsibilities include Radiation protection procedures associated with training and supervising health physics technicians, bioassay, instrument design, health physics sur- planning and scheduling monitoring and surveil- veys, waste disposal of contaminated sodium, and lance services, scheduling technicians to around- fire protection engineering are also different at an the-clock shift coverage as required, and LMFBR. maintaining current data files on radiation and contamination levels, personal exposures, and Differences in the operational health physics pro- work restrictions. They ensure that operations are grams at LMFBR facilities are discussed below with carried out within the provisions of the appropriate respect to ALARA, shielding, radiological source radiological hygiene standards and procedures; terms, waste management, process and effluent they establish an ALARA program to maintain the monitoring, sodium impurity monitoring, and occupational radiation doses as low as reasonably training. achievable; and they provide assistance and advice to the plant manager during radiological emergen- ALARA Programs. The primary responsibility of cies. These responsibilities are implemented by an operational health physics management is to reduce operational health physics program. the radiation dose to workers to as low as is

2 practicable, commensurate with sound economics e Authority from upper management for and operating practices. Paragraph 20.1 (c) of implementing an ALARA program with 10 CFR 20, “Standards for Protection Against the organizational freedom to ensure i Radiation,” states, in part, that NRC licensees d eve1o pm en t and imp1emeni:a t i o n o f should make every reasonable effort to maintain ALARA policies radiation exposure as far below the limits specified in that part as practicable. NRC Regulatory e Responsibility to ensure that the ALARA Guide 8.8 (RG 8.8), “Information Relevant to program incorporates ALARA manage- Maintaining Occupational Radiation Exposure as ment philosophy and regulatory require- Low as is Reasonably Achievable (Nuclear Power ments, with specific goals and objectives Reactors),” is addressed to applicants for a license for their implementation and tells them what information is relevant and should be included in their application with respect 0 Responsibility to ensure that the ALARA to keeping occupational radiation exposure to program has an effective measurcmment sys- employees to as low as is reasonably achievable tem, periodically reviewing the measure- (ALARA). RG 8.8 applies only to nuclear power ment system’s results and determining the reactors. RG 8.10 describes to all specific licensees degree of the system’s success. the general operating philosophy, which is the necessary basis for a program of maintaining occu- e Authority for producing procedures and pational exposures ALARA. practices by which specific ALARA goals and objectives will be achieved. Operational health physics management has established formal programs at nuclear generating 0 Authority to obtain resources needed to stations to ensure that occupational radiation expo- achieve goals and objectives necessary to sures to employees are kept ALARA. An ALARA maintain ALARA occupational radiation program has full management commitment to the exposures overall objectives of ALARA. It publishes for the design and operating group’s specific administra- e Responsibility to ensure that the ALARA tive documents and procedures that emphasize the program is implemented from initial plan- importance of ALARA throughout the design, ning through decommissioning of the b testing, start-up, operation, and maintenance plant phases of the reactor. The program continually appraises management of radiological conditions e Responsibility to review plant design fea- in the operating reactor by its on-site health physics tures, operating procedures, and mainte- staff. The effectiveness of the ALARA program is nance practices, and to audit the on-site reviewed and appraised at a corporate ALARA radiation control program, to ensure that committee consisting of representatives from the the objectives of the ALARA pro,,oram are design, operations, and radiation protection attained groups. The committee consists of key management and technical staff who have extensive back- 0 Responsibility to ensure that experience grounds in reactor radiation control, including gained during the operation of nuclear such areas as reactor layout, shielding, personnel power plants relative to in-plant radiation access control, ventilation, waste management, control is factored into revisions of operat- area and personnel monitoring, reactor operations, ing procedures, where necessary to ensure and reactor maintenance. The committee periodi- that procedures indeed do meet the objec- cally evaluates the overall ALARA program by tives of the ALARA program . assessing trends in occupational exposures or other radiation control problems, by _reviewing.reactor e Responsibility to evaluate trends: in the operating reports and radiation exposure profiles, exposure of station personnel, and if and by conducting on-site audits of the ALARA found adverse to take action to correct effort. them.

Findings of the ALARA committee are promptly Specific authority and responsibilities of the reported to top management, with appropriate ALARA committee include the following: recommendations for improvement or correction.

3 Responsibilities of radiation protection manage- 0 Reviews by experienced health physicists to ment with respect to the ALARA program include ensure the use of applicable current the following: LMFBR information.

Personnel are made aware of manage- Plant management is responsible for development’s commitment, their group’s ing plant radiation exposure objectives for specific ALARA goals, and their personal respon- functions and systems. Allocations are developed sibilities for estimated radiation exposure, which are based on the consideration of the total staff required to A well supervised radiation protection pro- operate and maintain the facility, and the radiation gram exists, with well defined responsibili- exposure objectives for individuals as well as the ties group. Objectives are also developed for radiation exposures of contract and utility personnel. Workers receive sufficient training After radiation exposure objectives, shielding Modification to equipment and mainte- criteria, radiation source terms, and time-access nance procedure is made when it will sub- requirements are identified, system designers pro- stantially reduce exposure at reasonable ceed with the objective to reduce total annual radia- cost tion exposure associated with the system to a level as low as reasonably achievable. Adequate staff exists to determine the location, operation, and job categories Management is assisted in its control of individ- associated with radiation exposure in the ual system design features that influence radiation facility exposure at an LMFBR by the performance of collective occupational radiation assessment. This Radiation protection staff examine ways to assessment is described in RG 8.19 for LWRs. The A. reduce exposure dose assessment process consists of the following evaluations: Adequate equipment and supplies for radiation protection is provided. . 0 Dose for specific categories of cells (pri- ALARA activities associated with the design, mary heat transfer cells, reactor contain- fabrication, construction, and preoperational test- ment building, reactor service building, ing activities involve the interaction of multiple etc.) engineering disciplines, such as radiation analysis, shielding, system, and component designers. Dose by skill classification (operators, Although radiation fields at LMFBR facilities may mechanical maintenance, electrical main- differ qualitatively and quantitatively from an tenance, etc.) LWR, the same analytical techniques can be applied. Health physics management establish Dose by system (auxiliary sodium, etc.) review and control procedures designed to incorporate and evaluate specific ALARA features. Review Dose by individual piece of major equip- and controls include the following: ment.

0 Established acceptable radiation exposure This information is compiled from input data levels provided by the following: the component, man- hours of operation and maintenance required for 0 Components and systems designed to each significant system component, frequency of achieve exposure and shielding objectives activity, and the cell or building number. This, together with the predicted cell dose rate forms the Appropriate documentation available to basis of radiation exposure studies. the ALARA committee so they can evaluate and manage the achieving of radiation ‘Radiation exposure information is periodically exposure objectives reviewed and updated as the system and component

4 design and analysis are developed. Using this sys- has been highly satisfactory, having extremely low tem, the significant contributors to radiation expo- levels of personnel exposure and only limited need sure can be identified, and appropriate ALARA for restrictions on access. As shown in Table 1, the action can be taken. The cognizant radiation pro- man-rem expenditures per MW/y(t) compares very tection engineers and the shielding designers from favorably with LWRs. each reactor manufacturer and the architect engineer participate in the reviews by approving the Dounreay-Radiation doses received by the shielding design of each component. All changes to staff at the Dounreay Nuclear Power De\.elopment plant design are reviewed and the impact on Establishment are extremely low, as indicated in ALARA determined. An ALARA program for Table 2. Doses are assessed by measurement of film normal operations, refueling, in-service inspection, badges, which are changed each month. Doses are and maintenance activities at LMFBRs can be compared with the film threshold and are seldom developed that fully meet the intent of NRC distinguishable from it. In no year has the total Regulatory Guides 8.8 and 8.10, and 10 CFR 20. dose above film threshold (film threshold is 0.01 rad per film), excluding deemed doses as described The review of operational health physics pro- below, amounted to more than 3.3 man-rem. Plant grams discerned no differences between LMFBR operations contributing most to this dose are and LWR facilities in their organization and admin- repairs to hot cell windows and component decon- istration of ALARA programs that can be credited tamination. For example, work on hot cell windows solely to the type of reactor involved. Evaluation of in 1980 contributed a total of 1.1 man-rem, or the ALARA programs at the five LMFBRs visited nearly 50% of the dose actually measured above in this investigation showed that differences of threshold. To achieve this low level, steel shield LMFBR organization and administration are more locks were placed above the fuel storage tank and a a function of licensing agency priorities and country of origin than of reactor type. Table 1. Man-rem expenditures per year [MWyltIl for LMFBRs and' LWRs The review found more noteworthy the differences between LMFBR and LWR facilities in their implementation of ALARA programs. During Reactor Type Man-Rem/MWy(t) both normal and off-normal operations during the lifetime of the plant, the LMFBR presents radio- BWR 0.92 logical hazards that stand apart from the LWR. PWR 0.5 Examples are potentially very large beta-gamma dose rate ratios during maintenance, tritium LMFBRs instead of noble gases as the primary source term to the environment, handling of contaminated Phenix 0.32 sodium, and training of plant personnel in sodium FFTF 0.99 technology. Specific differences are discussed DFR 0.45 throughout the rest of this Section.

Experience at Operating LMFBRs. Experience with radiation and contamination control at LMFBRs Table 2. Man-rem expenditures at the Dounreay Nuclear Power Development Establishment

Year -

-1980 -1981 ___1982

Average rem/year for radiation workers at the reactor facility 0.23 0.15 0.2.2

Total man-rem at the establishment 83.0 59.0 77. r3

5 portable lead shield was suspended behind the win- A single-barrier contamination control zone is dow being maintained. As another example, the established on the reactor top when access is gained total dose was 0.3 man-rem when maintenance during shutdown to the primary circuit. No signifi- work was done on the fuel charge machine. This cant contamination problems have arisen. Experi- does not include preparatory decontamination, ence in other areas of the plant has been similar. which was estimated to have contributed a further Only on few occasions has it been necessary to set dose no greater than 0.3 man-rem. up an active area with a double-barrier enclosure.

Monitoring the air in the building that contains The total registered dose annually, although low the reactor and irradiated fuel storage cell has in comparison with that for other types of nuclear shown negligible airborne activity. At the end of reactors, is dominated by components that are each year, the filters from the air monitors are ana- attributed rather than actually measured. lyzed using gamma spectrometry. Although evi- “Deemed” doses are attributed at the maximum dence of Co-60, Mn-54, and Cs-137 has been rate allowable (0.42 rem/month) per every film lost found, total activity is equivalent to the exposure of or damaged, and is assumed that every member of a single filter to one maximum permissible concen- the workforce receives the film threshold dose tration (MPC) for 1 hour. Tritium levels are rou- (0.01 rem). The threshold dose is the minimum tinely monitored throughout the plant and are well detectable dose that can be measured with the film below allowable concentrations. Samples of boiler badge. Clearly, these aspects become even more water average one hundredth of the level allowed in important as doses are reduced. Attention is being drinking water. Airborne tritium levels (from steam given to ways of reducing the attributed compo- leaks) are less than one thousandth of the derived nents, so that the actual dose can be more precisely air concentration allowed by regulations. quantified. Discharges from the plant stack have been Shielding of the reactor top has proven very sat- limited to noble gases and are indistinguishable isfactory with general levels of radiation in from background. Estimates of gas blanket leak- microrad per hour range at full power. Some addi- age, together with measured concentrations of tions have been made to the shielding in localized xenon-I33 and xenon-135, suggest the release of areas, and consequently there are now only a very not more than three curies in a year. No other spe- few places where radiation fields are even a few tens cies, such as 1-131, or long-lived particulates has of a millirad per hour. The secondary sodium pipe- been detected by the stack monitor. Tritium is work, which links the intermediate heat exchangers released in gaseous form from the gas blankets in the reactor tank with the steam generators in an above sodium surfaces and from decontamination adjacent building, is unshielded. In localized areas facilities. The total release is estimated to amount of low occupancy, the radiation level from this to about 150 curies annually. pipework is up to 3 millirad per hour. Radiation levels from fuel and active handling hardware are FnF-Plant personnel radiation exposures at similarly low. FFTF are low, as shown in Table 3. The values

Table 3. CY-82 personnel exposure summary

3rd Quarter 1st Quarter - 2nd Ouarter

Number in mrem/ Number in mrem/ Number in mrem/ Group Group Person Group Person Group Person

Operations 105 27 99 9 105 4 Maintenance 143 16 138 16 134 2 IEM Cell 22 4 23 4 33 1 Others 65 16 16 7 97 . 6 n

6 shown are within the statistical variation in the levels less than 0.5 mrem/h. A maximum reading dosimetry system for very low doses, and are at pre- of 1 mrem/h was obtained in contact with a pipe operational levels. Maximum individual dose for a containing primary sodium. quarter was 220 mrem and was associated with maintenance personnel working at another reactor An entry into cell 490 on June 26, 1982, facility during the second quarter. The highest showed general area levels less than 1 mrem/h. The exposure for personnel not associated with work at maximum loose contamination level in cell 490 was another reactor facility was less than 50 mrem for a about 5,000 dpm/100 cm2. Radiation and con- quarter. tamination levels associated with the IEM cell, the sodium removal system, and the radioactive liquid General area radiation levels have been very waste system were low, less than a few mn:m/h, and low. Dose rates measured in routinely occupied less than about 10,000 dpm/100 cm2 removable areas were less than the 0.2 mrem/h designed contamination. value. The few exceptions to design levels involved low-level, small area streaming through shield pen- EBR-//-Table 5 summarizes the radiation etrations. For example, the maximum shield con- exposure for EBR-I1 during 1970 through 1977. tact reading was 8 mrem/h on a shield located Data are provided for personnel in four major cate- about 10 ft above floor level. Other streaming loca- gories: reactor operations, instrumentation and tions are equally inaccessible or in areas with control maintenance, reactor systems mechanical extremely low occupancy. maintenance, and health physics.

Table 4 summarizes routine radiation survey The reactor operations group is coniposed of data. The values were extracted from weekly rou- four crews that man the continuous operation of tine survey reports and do not reflect special situa- the EBR-11. Each crew consists of 14 people, three tions, such as entry into cells containing primary of which are management. The remainder of reac- sodium and maintenance on radioactive systems. tor operations indicated in Table 5 is made up of Special radiation surveys were performed for these dayshift management and professionals in direct operations and the data are discussed below. technical support of EBR-11.

Routine radioactive contamination surveys are The operating crews have 11 technicians who also conducted. No contamination was detected in perform water and sodium sampling, control room the routinely occupied areas during the first opera- operation, and data recording for the power plant, tional period (cycles 1A and 1B) at FFTF. Local reactor plant, and sodium boiler building. Each of contamination control areas exist for specific tasks the data recording areas have general radiation performed under radiation work procedures. background levels of less than 5 mR/h, c,xcept for the sodium boiler building, which has an area Radiation surveys conducted in support of radiation field of about 20 mR/h and requires radiation work in the reactor show direct levels about 10 minutes occupancy per shift by one data ranging from background to about 1 rem/h. Con- recorder. The conditions stated above art for nor- tamination levels ranged from background to mal full-power operation; radiation levels are lower about 60,000 dpm/100 cm2. The 1 rem/h direct during shutdown periods. radiation was associated with primary sodium cover gas sampling evolutions during cycle 1A. The The instrumentation and control system main- radiation level was at contact with the gas tag sam- tenance group provides two technicians to work on ple trap (GTST) prior to installation of the shipping each of the four operating crews for routine instru- cover. With the shipping cover'in place, the contact ment maintenance and calibration. The remainder radiation level was 80 mrem/h, and personal expo- of the group of about 25 are management, profes- sure during the sampling evolution was limited to sionals, and dayshift personnel who sup'port the about 5 mrem. The 60,000 dpm/100 cm2 contami- experiments program and plant modification. nation level was associated with fuel handling, and Their work area radiation levels vary, depending on was a direct reading on a sodium contaminated plant condition, work routine, plant modification, floor valve. or experiment support.

An initial entry into cell 489 on July 8, 1982 The reactor crew was staffed by six or less per- (following cycle 1A), showed general area radiation sonnel during the report period, 1970-1977, and is

7 Table 4. Summary of routine radiation survey data during FFTF Cycle 1 n

Dose Rate (mrem/h) Dose Rate General Area Maximum Maximum Dependent Average Local During Local On On Reactor Location During Cycle Cycle 10/30/82 Power Comments

I. RSB, 500, Radioactive 0.5 4 0.5 No 4 mrem/h due to temporary storage of IEM Cell components.

2. RSB, 550, General Area 0.2 1 1 No 1 mrem/h due to floor valve adapter. 3. RSB, 520, Rm. 204 0.2 0.2 0.2 No Looking for potential leaking to secondary chilled water from IEM Cell Removal System.

4. RSB, 520, General Area 0.2 0.2 0.2 No Local dose could be dependent on cladding breach & release of fission gasses.

5. RSB, 543, General Area 0.2 0.2 0.2 No Same comment as for 4. 6. RSB, 533, General Area 0.2 4 0.2 No 4 mrem/h due to small leak of 41Ar from RAPS instrument line.

7. HTS-S, 550, General Area 0.2 0.2 0.2 No 8. HTS-S, Room 481, 493, 0.2 2 2 Yes 2 mrem/h due to valve stem penetration and and 482 482(P-622)into Cell 489, reactor at 100% Source Na.

9. HTS-S, 550, Room 485 0.2 3 2 Yes 2 & 3 mrem/h due to 24Na radiation streaming around shield plug to Cell 492-E, reactor at 100%.

IO. RCB, 580 Mezzanine 0.2 3 1 Yes 1 & 3 mrem/h due to contact measurement on low level flux monitor cooling system. Sourc is transported activation products and I'N.

1 I. RCB, 550, General Area 0.2 0.5 0.2 No General area dose rates in RCB at full power are not significantly different from background. 0.5 mrem/h in Primary Sodium Pump 1 Pit.

12. RCB, Head Compartment 0.2 6 6 No 6 mrem/h due to depleted uranium shielding in two small local areas. General area 0.2 mrem/h.

13. RCB, 520, Room 540 0.2 0.2 0.2 No 14. RCB, Stairway 517 West 0.2 0.6 0.6 Yes 1s. RCB, 520 Room 564 0.2 0.5 0.5 Yes 0.5 mrem.h at locked barrier to restricted access radiation zone.

16. RCB, 520, Room 5548 0.2 0.2 0.2 No 17. RCB, 500, Room 562 0.2 0.2 0.2 No

18. RCB, 500, Room 570 0.2 9 9 Yes 9 mre/h due to 24Na radiation streaming around shield plug to Cell 567 (EM pump cell), reactor at 100% power.

19. RCB, 520, Room 5538 0.2 0.2 0.2 No 20. RCB, 500, Room 565 0.2 0.2 0.2 No 21. RCB, 520, Room 5368 0.2 8 6 Yes 8 mrem/h due to 24Na radiation streaming through penetrationin a small, localized area about 10 feet above floor.

22. RCB, 500, Room 561 0.2 0.5 0.5 Yes 0.5 mrem/h due to 24Na radiation streaming around shield door to 532C.

23. RCB, 520, Room 522B 0.2 0.5 0.5 Yes 5 mem/h at locked gate to 552A2,source is "Na radiation streaming into 552A from HTS (521) n Table 5. Exposure histories for EBR-II work groups

Individual Individual Individual Number High LOW of Exposure Exposure Total Average People (mrem) (mrem) (man-rem) (mrem)

Rx Operations

1970 51 200 1.100 21 1971 48 440 3.210 66 1972 52 600 2.600 50 1973 60 320 3.275 55 1974 58 495 5.435 92 1975 65 475 6.675 103 1976 67 455 6.953 104 1977 69 510 10.047 183

I&C Instrumentation

1970 13 500 1.810 139 1971 18 740 4.120 228 1972 18 765 6.150 341 1973 14 350 2.290 163 1974 26 1330 6.840 263 1975 31 430 4.470 144 1976 34 625 5.980 176 b 1977 33 543 3.982 121

Rx Crew E Maintenance

1970 5 1095 250 2.955 591 1971 5 1420 315 4.500 900 1972 4 965 230 2.400 600 1973 3 715 75 1.660 315 1974 5 715 75 1.660 332 1975 6 1595 75 6.070 101 1 1976 6 1210 120 5.160 860 1977 6 1224 667 5.398 900

Health Physics

1970 4 1030 140 2.280 570 1971 5 940 300 2.965 593 1972 5 1065 485 3.950 790 1973 5 670 380 2.300 460 1974 7 655 210 2.420 345 1975 7 805 160 3.585 512 1976 7 730 260 3.735 535 1977 8 765 25 1 3.749 468

9 the highest exposure group. The group inserts and The average exposure and the total man- removes all highly radioactive experiments, and rem were significantly higher in 1977 than performs the maintenance of fuel handling equip- previous years for reactor operations per- ment. The group also handles deposition of sonnel irradiation test hardware. The INC instrumentation group exposures The health physics group provides the necessary have been reasonably steady, with a signifi- radiological controls for routine operation and for cant decrease in 1977 the protection of personnel while they are performing their tasks. This is accomplished by four techni- The reactor crew showed an increase in cians on day shift and one technician on each individual low exposure, with other items off-shift. being reasonably steady for the last 3 years

Health Physics technicians normally perform The Health Physics group shows few fluc- routine surveillance, conduct radiation level mea- tuations in radiation exposure surement, and perform surveys to assess contamination. The day shift technicians also support the The highest individual exposure in 1977 mechanical and electrical maintenance activities, was 1224 mrem, compared to the U.S. and experimental activities as required. DOE limit of 5000 mrem.

Table 6 summarizes the exposure required to operate EBR-I1 during the report period, 1970 Shielding through 1977. This includes the number of exposure personnel, the individual high exposure, and The ALARA criteria of NRC RG 8.8 are important the average exposure of the group. There has been a to containing personnel radiation exposure to very steady increase in man-rem from 1974 through low !evels. The guide explains that merely control- 1977. From Tables 5 and 6, the following can be ling the maximum dose to individuals is not noted: sufficient; the collective dose to all reactor plant

Table 6. EBR-II annual exposure summaries

Individual High Exposure Total Average -Year Number of People (mrem) (man-rem) (mrem)

1970 73 095 8.145 112

1971 76 420 14.795 195

1972 79 065 15.100 191

1973 82 670 8.81 107

1974 96 330 16.355 170

1975 09 595 20.800 191

1976 14 210 21.828 191

1977 16 224 23.176 200

10 --

personnel must also be kept ALARA. This is an that required functions are safely provided important point; the total man-rem collective dose throughout the lifetime of the pliint could increase because of inefficiencies involved in having to use a large number of individuals to To limit the radiation exposure of operat- accomplish such tasks as maintenance and repairs. ing personnel and the general public to RG 8.8 mentions numerous exposure reduction that less than required by applicable sec- techniques that can be used to lower dose rates and tions of 10 CFR 20, 50, and 100. occupational exposures. Many of the techniques are concerned with minimizing the radiological The thickness of shield walls are set by both source terms. Where this is not possible, shielding shielding and structural requirements. 'Walls are is an effective way to reduce the effect the radiologi- usually constructed of ordinary concrete. LMFBR cal source term has on occupational exposure. concrete shield walls can be errected according to the standards used for LWRs (see Regulatcry Guide Shielding is designed to provide radiation protec- 1.69 and ANSI N101.6). In some local areas, other tion to operating personnel, the general public, and shielding materials, such as steel, are used. Shield plant equipment. Shield design must perform a walls constructed of block have usually been variety of functions to provide this radiological avoided, for seismic considerations. Shield walls or protection under normal operating conditions. portions of shield walls subject to removal to per- These functions are as follows: mit access for equipment repair or replacement are designed to be knocked out and replaced, or pro- To permit operating personnel access to vided with removable access plugs. In ordw to min- required portions of the plant in order that imize the number of areas in the plant with high they may operate the reactor and its associ- radiation levels, cells that contain significant radio- ated systems at rated power and normal active sources are grouped. capacity (including routine inspections and maintenance of accessible equipment) Shielding calculations for determining wall thickness can be performed by a computer code To permit personnel to refuel the reactor with a detailed geometric model that represents the (including those functions required to pre- significant three-dimensional aspects of the sources pare new fuel for reactor service and ship and shields. Codes used in the shielding arialysis of spent fuel for reprocessing or storage) LWRs are adequate for LMFBRs. Outside the reactor cavity, the sources present are activated sodium, fission, and corrosion products. To permit access within a reasonable time after reactor shutdown to the high radia- Computer programs employed in the d!esign of tion portions of the restricted area (which shield walls use the discrete ordnance transport or are maintained as exclusion areas during point kernel integration techniques to analyze the normal operation) bulk shielding provided by the walls. In areas of the shield containing discontinuities, such as ducts and 0 To limit the neutron activation of second- penetrations, the design methods employ transport ary and intermediate sodium, such that the and point kernel scattering methods to provide induced radiation dose rates from this shielding equivalent to the surrounding bulk shield- radioactive source will not require estab- ing in the vicinity of the duct or penetration. lishing restricted areas Shield walls with locked access doors arc: used to To maintain as continuous access areas dur- isolate high radiation areas of the plant. In addi- ing normal operations all areas of the site tion, these areas are equipped with audible or visi- outside of the reactor containment building, ble alarms that signal to a control point that a high reactor service buildings, cells in the plant radiation area is being entered. All locked radiation service buildings, and any secondary sodium areas are equipped such that no individual can be piping penetration cells at the intermediate/ prevented from leaving them. reactor containment building interface Access to cells containing radioactive.-bearing To protect structural components, equip- components are provided through the wall!; or ceil- ment, and nuclear instruments in order ing. The effectiveness of the shielding in meeting

11 the design dose rates external to the cells is main- tions of experiments used either the bare beam or sim- tained by use of shield doors or shield plugs, which, ulations of calculated neutron energy spectrum if made of concrete have the same thickness as the incident on the specific portion of the LMFBR shield shield wall penetrated. If other shielding materials system, for example, at the reactor core-blanket inter- are used, the design thickness provides shielding face. The design of spectrum modifiers to simulate equivalent to that of the wall. Steps and off-sets are incident spectrum have been used extensively in TSF provided in all shielding doors and plugs to mini- programs. mize radiation streaming. For areas requiring multiple penetrations, shielded cavities and chaseways Experiments related to CRBR shield design and are provided. conducted in the FFTF program on the TSF were (a) mock-up experiments of the lower shield region Shield performance and design, and basic of the fuel assemblies, which included representa- nuclear data, are being verified to support LMFBR tion of the streaming paths between assemblies and systems. A series of experimental configurations through-shield orifice holes, (b) mock-up experi- simulating portions of LMFBR shield systems have ments of the reactor vessel closure assembly pene- been designed and experiments have been con- tration gaps, which included parametric ducted. The principal experimental program has investigation of offsets and gap widths for annual been conducted at Oak Ridge National Laboratory. gaps and slits in iron, (c) deep-penetration neutron A number of other experiments related to shielding attenuation experiments in sodium, carbon steel, have been conducted in the Argonne National Lab- and stainless steel, (d) mock-up experiments to oratory and support the Clinch River Breeder Reac- define neutron streaming in the primary inlet and tor Project (CRBRP) nuclear design. The latter outlet piping penetrations of the reactor cavity program, which provides data useful in evaluating wall, the isolation valve cell, and the shield walls neutron flux and gamma flux distributions in adjacent to the heat transport system (HTS) cells, LMFBRs, is described below. and (e) mock-up experiments of the reactor closure head and head compartment shield system to In general, three types of experiments have been define neutron and gamma shield effectiveness. conducted to support shielding design: (a) experi- These experiments employed spectrum modifiers to ments to measure the basic neutron and gamma atten- obtain prototypic neutron spectra incident on uation characteristics of candidate materials for use in shield systems. The CRBR shielding program has LMFBR shields, (b) experiments to parametrically expanded from this original FFTF program to define shield configuration and dimensional effects include a number of experiments simulating the on shielding performance, and (c) prototypic or sim- material and laminar configurations of tentative ulation experiments of portions of LMFBR shield LMFBR radial, upper axial, and lower axial systems. In addition to these three types of experi- shields, and provide tests of calculation methods in ments, measurements of neutron cross section data in prototypic shield configurations. Table 7 describes the candidate LMFBR shield materials have been the scope (completed and planned) of program conducted. The LMFBR shielding experiment pro- experiments to support LMFBR shielding design. gram has progressed from experiments designed to support the FFTF system, to experiments to support the CRBRP system. The scope of experimental pro- In addition to the experiments conducted at the grams conducted on FFTF shielding have provided a ORNL power shielding facility, an extensive pro- sound base of experimental data for method and gram of experimental analysis has been conducted nuclear data verification. FFTF experiments were by ONRL with the reactor manufacturers. These conducted from 1969 through 1973; LMFBR shield- programs were employed to test methodology and ing experiments were initiated in 1973 with emphasis basic nuclear data in terms of design solutions of on CRBRP shielding beginning in 1974.1-12 LMFBR shields. The purpose of experimental analysis program was manifold in that it has pro- The ORNL shielding experiment program was con- vided (a) guidance in experiment planning, to ducted at the ORNL power shielding facility. Experi- ensure adequate data for testing methods and basic mental configurations employed a collimated beam nuclear data, (b) guidance in planning the develop- of neutrons from the TSR reactor configuration as an ment of design methods required to design shield- approximate planar source of neutrons incident on ing for LMFBRs, (c) reactor manufacturers with slab simulations of LMFBR shield systems. Simula- the vehicle to test and use the methodology prior n

12 Table 7. Experimental program related to CRBRP shielding I\

Experiment Sponsora Objective

Radial blanket and LMFBR Provide experimental verification of shield simulation neutron flux attenuation through radial blanket and radial shield simulationa

Near core gamma CRBR Provide experimental verification of gamma heating rate distributions in (1) stainless steel and Inconel radi a1 shield configurations and (2) at core- blanket and blanket-shield interfaces

Upper axial shield CRBR Provide experimental verification of neutron attenuation methods based on the CRBR reactor closure head siinula- tion with sodium pool simulation

Reactor vessel support CRBR Provide experimental verification of area methods used to analyze reactor vessel support area neutron streaming

Concrete rebar CRBR Provide experimental verification of neutron flux attenuation and secondary gamma production of reinforced concrete

LLFM Provide experimental verification of ex- vessel LLFM neutron flux and count rate

Sodium pipe chaseway FFTF, Provide experimental verification of CRBR neutron streaming in sodium pipe chaseway

Effect of shield LMFBR Provide experimental determination of heterogeneities- streaming effects due to compositimon upper axial and heterogeneities (Upper Axial and Radial radial shields Shields)

Duct streaming- LMFBR Provide experimental determination of lower axial shield streaming effects due to composition heterogeneities (Lower Axial Shield)

Annular gap and FFTF Measure neutron streaming through iron slit streaming shields with annular gaps or slits simulating closure head shielding.

Core assembly FFTF Provide experimental verification cf shield streaming streaming through coolant holes of' core shields.

13 Table 7. (continued) n

Experiment Sponsora Objective

Sodium penetration FFTF Measure shielding effectiveness of sodium pool attenuation through up to 15 feet of sodium.

Inconel penetration FFTF Measure neutron shielding effectiveness of Inconel shielding material.

Gamma ray experiment FFTF Provide experimental verification of neutron transmission through 34 inches of carbon and provide verification of transmitted gamma spectra and dose. a. Sponsor of experimental involved in providing guidance to AEC-RRD for the specific project. LMFBR experiments are through efforts of reactor manufacturers. to, or in parallel with, design analyses, and (d) cor- long plant shutdowns, neither of which is economi- relation of the design analysis of portions of the cally favorable. The corrosion product transport CRBR shield system in order to experiment and (or “crud transport” problem as it is called in water provide realistic estimates of the uncertainties and reactor technology) is well known, and considera- experimental-to-calculated bias associated with ble maintenance experience has been gained on shield performance. The last item provides a sound both water and sodium-cooled reactors. basis for shield design in that the detailed design support calculations performed by reactor manu- Calculations made for reactor plants operating facturers and the independent support calculations with long fuel cycles and at high core outlet temper- performed by ORNL used the techniques and atures (600°C) predict radiation levels above 1 R/h methods demonstrated in the experiment analysis adjacent to the rimary system after decay of the effort. sodium 24. 14-lg

Results from the LMFBR experiment and experi- Present reactors, with lower outlet temperatures ment analyses programs have demonstrated the and shorter fuel cycles, substantially reduce corro- validity of the analytical techniques and shield sion product transport, to produce radiation levels design methods. Detailed results and comparisons of about 100 mR/h is being observed. Extrapola- of these structures are documented in References 3 tion of calculations to the existing reactor situation through 13. results in reasonable agreement 17.

Radiological Source Terms. In an LMFBR, clad- Reactor experience and test loop data show that ding and in-vessel structural materials become both the release of irradiated material and its distri- radioactive from the neutron flux. Radioactive spe- bution within the circuit are nonuniform. The most cies build up in the primary circuit external to the abundant corrosion product in LMFBR nuclides is core, caused by neutron activated material corrod- Mn-54. CO-60and Co-58 are also produced but are ing out of the piping walls and subsequently being not released at as high a rate as Mn-54, nor are they transported in the primary liquid sodium to the pri- transported throughout the circuit to the same mary components (for example, pumps, heat degree. A major fraction of the deposited corrosion exchangers, valves, and sodium purification sys- product activity adheres to the deposition site and tems). The deposition of these radioactive species is not removed by sodium draining or sodium gives rise to gamma radiation fields intense enough removal processes. to cause difficulties in maintenance and repair operations. Moreover, such repairs have required No one control method has been shown suffi- large amounts of manpower and have resulted in cient to solve all of the activated corrosion product n

14 - ...... - ...... -

transport problem; a combination of methods is were all ineffective at removing deposited radioactivity; required. Methods under investigation have to do the acid etch was somewhat more effective than the with the effect of oxygen concentration in the other At the EBR-I1 reactor, depos- sodium, and sodium temperature on nuclide ited Mn-54 and Co-60 have been observed on the pump release, development of traps, improved fuel clad- and heat exchanger and in the cold trap. E:xamination ding alloys, and removing deposited activity of 316 SS fuel cladding from a high temperature fuel (decontamination). Trapping and decontamination test showed substantial preferential corrosicin of Mn-54 show promising early results. Reduction of the oxy- and preferential retention of (20-58 and Co-60 at gen level of the sodium to correspond with the 650”C, but very little at 510”C.25-26 115°C cold trap temperature is only partially effective at reducing Mn-54 and Co-60 release. The radiation levels observed in the plants above Table 8 presents information concerning the sig- are in reasonable agreement by a factor 01’ three with nificant radionuclides and activated corrosion those predicted using the techniques of References 14 product transport. The most prevalent nuclides and 27. A significant result is that the contribution of seen in LMFBRs are Mn-54 and Co-60; others are “nuclear corrosion” (recoil ejection plus eist neutron observed in smaller quantities. It can be seen from sputtering) has been shown unimportant, in apparent disagreement with earlier hypothesis.28 Using the the Table that little can be done to eliminate Mn-54, whereas some control over (20-60 formation is pos- techniques of References 14 and 27, it has been calcu- sible by restricting the cobalt content of fuel clad- lated that a uniform removal rate of 25 p/yr by fast ding and ducts to 0.05‘70, and by restricting the use neutron sputtering will lead to primary system radia- of cobalt-based bearing and hard-facing materials. tion levels well in excess of 10 R/h, which is contrary to actual observations in operating reactors. LMFBR Experience. At the Dounreay fast reactor, repair of a leak that took place in a radiation field Before operational data on LMFBRs became avail- equal to that predicted for high-temperature LMFBR able, investigators in the United States, Germany, operation required several hundred people and several Japan, and the United Kingdom began nuclide months downtime for repair. At the BR-5 LMFBR release and deposition studies in small test loop sys- in the USSR, fission products, Mn-54, and Co-60 were tems. The need for mass transport data, in addition to observed on piping. Treatment of the piping with acid that generated from the more conventional corrosion etches was necessary to removal all of the deposited studies of steel and sodium, has arisen for two rea- radioactivity. 19-20 At the BOR-60 reactor in the sons: (a) it can be shown that an average m,aterialcor- USSR, radiation levels of up to 100 mR/h adjacent to rosion rate of 5- 10 p/yr for fuel cladding (completely primary system piping was observed after 3 years of satisfactory with a wall-thinning standpoint) will give operation. The core outlet temperature varied between rise to intolerable radiation levels near the primary 510 and 550°C. The Co-60 and Mn-54 radioactivity system components, and (b) the problem nuclides was spread throughout the circuit with greater nuclide (Mn and Co) are present as only minor constituents deposition in the cold leg.21 At the KNK-I1 reactor in of the cladding alloys. Existing corrosion theories do Germany, radiation levels of 100 mR/h were measured not treat minor alloying element behavior in suffic- near the heat exchangers after 146 days of full power ient detail. operation. The core outlet temperature was 550°C. The radiation level was reduced only about 15% by Most corrosion and deposition studies (such as draining the sodium. The principal nuclides were found in References 25, 29-34) investigated the release Mn54, Co-60, Zn-65, and Ag-1 1Om. Mn-54 tended to and deposition of Co-60, Mn-54, Fe-59, Cr-51, and preferentially deposit on a nickel rich surface in the hot Co-50, at temperatures between 550°C arid 600°C. leg but not the cold leg.22 It is noteworthy that consid- The oxygen level in sodium was controlled by cold trap- erable Ta-182 was produced in the KNK-I1 core, but ping and measured by a variety of methods: cold trap only small amounts were detected in the piping. At the temperatures were in the range of 115- 160°C. These Rapsodie reactor in France, radioactive in Mn-54 was studies agreed with reactor observations and with each observed throughout the piping system, particularly in other; the most significant similarities and differences the cold leg. Very little Co-60 was observed. Radiation are as follows: levels up to 200 mR/h were observed on the pump and piping. Washing with water and alcohol, and treatment Mn-54 was the most abundanii nuclide with a relatively mild nitric-phosphoric acid solution reduced from irradiated stainless steel; Cr-5 1,

15 Table 8. Corrosion product radionuclides n

Formation Half-Life Gamma-Ray Nu c1 id e Reactions(s) (days) Energy, MeV Comments

Mn-54 Fe-54(n,~)~ 313 0.84 The most prevalent nuclide. Mn-55(n,2n) Eliminating all the others still leaves a significant problem.

CO-60 Co-59(n,y) 1913 1.17, 1.33 Source is Co impurity in Ni-60(n, p) nickel and Co-base wear pads, bearings, and hardfacing materials.

CO-58 71 0.81 Use of a ferritic steel for fuel cladding eliminates most CO-58.

Fe-59 45 1.10, 1.29

Ta- 182 115 1.12, 1.11 Source concentration increases in Nb-bearing steel is used in the neutron flux.

Zn-65 243 1.11 Sorce is Zn in sodium or from contamination by ZnCr03 rust-proofing paint.

Ag-l10m 253 0.65, 0.76 Suspected source is Ag 1.47 impurity in nickel. This nuclide has not been observed in the significant quantities in the U.S.

Cr-5 1 28 0.32 Although substantial, low gamma energy makes Cr-5 1 transport inconsequential. a. Dominant reaction.

Fe-59, Co-60, and Co-58 were released in pro- more heavily in the lower temperature por- portionally lesser amounts. However, the high tion of the loops. CO-60 and Cr-51 showed concentrations, long half-life, and high preferential deposition in the loop hot legs. energy gamma radiation of Co-60 make it a Mn-54 deposited in loop hot legs in the first more important problem than Cr-51, Co-58, few hundred hours of loop operation, and or Fe-59, which have been observed only in gradually migrated to the lower temperature small amounts in actual reactor operation. regions.

For the most part, Mn-54 was transported The loop cold traps were more or less inef- throughout the test circuits and deposited fective in removing radioactivity at rates

16 more than proportional to the cold trap frac- in the case of CO-60 deposition patterns. tional flow. They did, however, act as irre- Proportionately more Co-60 is reported in versible sinks for whatever cobalt and reactor cold legs and cold traps than in the manganese deposited in them. The concen- test loops. Also, examination of fuel pins at tration of Mn-54 in cold traps apparently the top of an EBR-I1 fuel assembly (0.5 m depends on how much Mn-54 has deposited above the fuel zone) showed no preferential upstream of the cold trap branch line. This deposition of Co-58 or Co-60 on the fuel pin suggests that to be most effective as a radio- surface. nuclide sink, the cold trap branch line should come from a reactor hot leg or above the Usually, radionuclide behavior is the same in loop core. and pool reactors. However, distinctive behaviors are sometimes observed, for example, the behavior of Reducing the Oxygen level by reducing particulates in sodium. This is important I,ecause trap temperature in an apparent sodium is the primary coolant for all of the reactors in the c0-60 Source rate but Only a except DFR, where it is NaK. DFR has another speci- small decrease in the Mn-54 source term. ality: a vented fuel element. This causes the fission Quantitative determination Of Source rates product inventory ofthe DFR coolant to be ve1.y high. have shown Mn-54 to be preferential1y The behavior of radionuclides in the DFR's primary and c0-60 to be system, therefore, cannot always be compared with retained from irradiated 3 16 SS speci- the results of other reactors. mens.25535

0 At all locations in the test loop, deposition Experience with radionuclides in operating reactors was enhanced at regions of increased fluid has had varying degrees of importance to health phys- turbulence, where the hydrodynamic ics in particular systems and at different times. As men- boundary layer thickness is reduced. The tioned, activated corrosion products are most significance of this data is that increased important for system contamination, in particular Mn- nuclide deposition can be expected on 54, Co-60, and Co-58. Fission products becom: more pumps, valves, and other locations where important after operation with defective fuel elements; additional turbulence is introduced. of these, next to (3-137, (3-134, and 1-131, the Unfortunately, maintenance problems are nuclides Zr/Nb-95, Ba/La-140, Ce-141 and Ce- 144 are expected to occur at these regions. On the of most concern. It has been reported that the ra d' ioac- other hand, the data can aid in the design tivities of the corrosion products Co-60 and bdn-54, of nuclide traps to enhance deposition of and of the fission products Ba/La-140 and Nb-'95, are the nuclides and prevent their spread into comparable to each other in reactor operations where the reactor primary system. 0.1 to 0.2% of the fuel elements are defective, artd that the deposited corrosion product activity has been con- Radioactive material was firmly deposited stant over three years.36 Published tables present the on the test system piping at temperatures radionuclides detected in reactor systems.37 Research above 400°C; sodium removal processes programs have investigated the behavior of racfonu- such as draining, water, steam, or alcohol clides in primary systems for ~0~-60,38339DFR,m rinse removed very little activity. Use of acid EBR-II,41 RAPSODIE and PHENIX,42 and KNK- solutions was necessary to remove deposited K43 Additional information is available on radionu- activity (described in References 20, 23, 24, clide transport for PHENIX4 and the effect (of the and 34). Significant penetration of activity special run beyond cladding breach (RBCB) program large than 10 p into the pipe wall has been on radionuclides in the primary system at EBR-II.45- 48 observed even at 400°C. Under certain conditions, deposits build up to a limiting thickness and then are sheared off.34 The Summarizing all corrosion and deposition experi- phenomenon could have practical implica- ence gained with operating LMFBRs, it can bc con- tions for removing deposits from an interme- cluded that, so far, activation product radionuclides diate heat exchanger in a pool reactor. in primary systems have not been a serious problem for operation. Extended downtime because of large Good agreement does not exist between test radiation doses has been reported only from D17R.49 /\ loop data and reactor operating experience However, knowledge of the behavior of radionuclides

17 in primary systems is very important in order to be radiated pin that had 10.6% burnup. Over 97% of prepared for difficulties arising from system contami- the released fuel was firmly bound to the SS surface nation. of the system and could only be removed with Q strong There are two major sources of fissile materials in an LMFBR reactor system. The first is surface The expression “tramp fuel” is often used in the contamination of new fuel elements during their nuclear industry and is defined as the amount of fabrication. Generally, g of fissile material per fuel in a reactor system in the core region responsi- cm2 of surface is found, the fissile material usually ble for the background level of fission products in reported in U-235 equivalents. The second source the primary system. Reported values are in the of fissile material is fuel released from defective range of a few mg U-235 equivalent. Such a value fuel elements. Even if the amount of fuel released can be explained by the surface contamination of cannot be predicted, and assuming very small fresh fuel elements mentioned above. Data from releases of fuel, after operation of a reactor with EBR-I1 indicate, however, that the tramp fuel is in defective fuel pins, the released fuel will dominate particulate form.55 The literature does not report other contamination. an increasing amount of tramp fuel in reactors. This is probably a result of improved fuel manage- U02, Pu02, and (U, Pu)02 react with sodium if ment procedures. the oxygen potential of the fuel or sodium is large Because the oxides themselves are stable in It is concluded from thermodynamic consider- clean sodium, transport of fissile materials by sodium ations and from observations of sodium systems will be either in the form of oxide particles, because of that the concentration of fissile materials in the cir- the extremely low solubility of the oxides, or of reac- culating sodium is very small. Strong deposition on tion products with sodium, such as Na3M04 (with the walls of the system and particulate transport to M = U or Pu). In any case, as with other oxide- the cold traps will occur. All available information forming nuclides, the fissile material concentration in from operating reactors are, however, very meager, .. sodium will be very low, and rapid deposition or pla- indicating that fuel in the primary system has not teout will occur.51 been a problem for operating reactors. Ventilation Observations of operating reactors confirm the plateout of fissile material. The concentration of Design Objectives. As they relate to radiation plutonium and uranium in the sodium at EBR-I1 protection, the design objectives of the heating, has always been below detection levels: less than ventilating, and air conditioning system (HVAC) 0.5 pCi/g sodium of Pu-239/240, and less than are the same for LMFBRs as LWRs during normal

1 ppb U.41 947 Similar low values were reported for operations, including anticipated operational RAPSODIE. However, higher concentrations were occurrences. The design objectives are as follows: found for this reactor after cladding failures. Val- ues of up to 200 pCi/g sodium for Pu-239/240 and To limit in-plant buildup of airborne radio- up to 2 ppm for uranium were found.52 There is a activity during normal plant operations so wide scattering of the measured data (a factor of that exposure to a person from radioactive lo), probably because the transport is primarily of materials in the air will not exceed the design particulates. Widely scattered data were again radiation dose rate for a specific habitable observed during investigations of the cold trap bas- zone; plant ventilation complies with the kets of DFR. It is estimated of DFR that the pri- requirements of 10 CFR 20, Appendix B, mary system contains less than 10 g uranium and Table 1 less than 10 mg plutonium.40 The special condition of vented fuel elements at DFR have to be kept To limit the migration of airborne radioac- in mind, however, when attempting to transfer tivity from areas of high radiation poten- these data to other reactor systems. tial to areas of low radiation potential

In an experiment that purposely used defective To ensure habitability of the main control fuel pins with artificial cladding defects of room during any postulated abnormal 30 mm2, about 4 mg of fuel were released to the condition, to enable control or shutdown of the plant sodium from a fresh pin and 1.4 g from a preir- n

18 To provide the capability to isolate the Another source of radioactive gases in an LMFBR reactor containment building atmosphere is reactor cover gas. Activation of impurities in the after an accidental release of radioactivity. sodium, and direct activation of Ar-40 to Ar-41, contribute to the activity in the cover gas. E,ven Ne-23 Design objectives, equipment, instrumentation, appears from an (n,p) reaction with Na-23, but its and methodologies for airborne radioactivity moni- half life is short (38 s). The main design requirements, toring at LMFBRs do not differ significantly from however, for shielding and ventilation control, is those at LWRs. Present fixed and mobile continuous based on permitting reactor operation with leakage air monitors are employed in conjunction with porta- occurring in a specified fraction of the fuel pins. For ble air sampling equipment to satisfy the require- FFTF, this fraction of defective pins was jet at 1Yo. ments of 10 CFR 20, and to verify that radioactive Calculated activity caused by fission gas in the reactor atmospheric contamination is as low as reasonably cover gas for this design basis condition is presented achievable. Monitoring is continuous in frequently in Table 9. Since failed fuel will likely never' approach occupied operating areas adjacent to potential radio- 1Yo, actual activities will likely be far below these lev- active sources, in frequently occupied areas including els. Table 10 lists the radioactive gaseous releases at radiation zones or cells that house numerous process ANL-W and EBR-11. The data in these Tables demon- system control panels, and in ventilation that serves strate that the concentration of the fission gas and the the reactor containment building or reactor service activation noble gases are comparable. building. Table 9. FFTF design basis cover gas Source Terms. Gaseous radionuclides produced activity in LMFBRs include the activation products, Ne-23, Ar-39, and Ar-41; the noble gases, xenon and krypton; and tritium. Activity Isotope (~i/m3) Fission product gases are produced during the fission process, both directly from fission and from the Xe-13 1ma 5.43 x 10-1 decay of radionuclides to stable gases, as in the decay Xe-133m 1.47 x IO1 of iodine to xenon. Most of the fission gas is xenon. Xe-133 2.67 x lo2 The second most abundant gas is krypton. The yield Xe-135 1.26 x lo3 of stable fission product gas, Le., the fraction of fission products that become stable gas atoms for a Kr-83m 6.82 x IO1 mixed U02- Pu02 fast reactor is about 0.27.56 Kr-85m 1.34 x lo2 Kr-85 9.30 10-9 Much of the fission gas produced in the unrestruc- Kr-87 1.80 x lo2 tured fuel, where temperatures are lowest, is retained Kr-88 2.64 x lo2 in the fuel grains during normal operation. In the equiaxed and columnar grain regions, most of the fission gas escapes to the central void region or through a. The m designates the metastable state. cracks to the fuel cladding interface.

The processes by which fission gas bubbles nucle- Because noble gases do not form chemical com- ate, grow, defuse, and eventually collect and condense pounds under LMFBR conditions, it should be possi- at grain boundaries are extremely complex. Below a ble to describe their behavior by models, involving temperature of about 1300 K, fission gas mobility is solubilities and diffusion coefficients in In very low and essentially no gas escapes. Between 1300 reality, however, simple equilibria are not obtained and 1900 K, atomic motion allows some diffusion to and more empirical models are required. Moreover, it take place, such that over a long period of time an is important to distinguish between fission gases dis- appreciable amount of gas can reach escape surfaces. solved in the sodium and those transported as bub- Above 1900 K, thermal gradients can, in days or bles. months, drive gas bubbles and pores over distances comparable to grain sizes. Gas release then occurs Cover gas bubbles are always contained in the when bubbles reach a crack or other surface con- sodium of reactor~.~~-~OFission gas butibles are nected directly with a free volume. released from failed fuel elements. It is generally

19 Table 10. ANL-WEST and EBR-II radioactive gaseous releases, 1968-1978

Total ANL-W Curies EBR-I1 Curiesa , __Year

1968 837 0.01 (0.001 070) 1969 130 0.0 1970 84 0.1 (0.1%) 1971 74 8.6 (12%) 1972 127 20 1973 803 674 (84%) 1974 666 515 (77%) 1975 669 482 (72%) 1976 556 379 (68%) 1977 635 461 (73%) 1978 297 143 (48%) a. EBR-I1 percentage of total ANL-W shown in parentheses. accepted that fission gas bubbles are transferred to the In LMFBRs, the last case is the most significant in cover gas within minutes, even after breakdown of the delay of dissolved gases to the cover gas. The delay bubbles into small sizes, as is assumed for EBR-IL61 is usually expressed in the form of a delay or degassing probability, Ad.65 For radioactive nuclides with a Tramp fuel and bare fuel surfaces in contact with decay constant for the radioactive decay, Xi, a simple sodium, as in the case of cladding failures, and in model for equilibrium conditions may be written: some case even “leakers,” will produce such low concentrations of fission gases that the fission gases will not be detected by the failed fuel detection systems. RF. = R./P. = ’d The delay in the release of dissolved gases to the cover 1 1 1 A6 + A. ’ 1 gas may be caused by several effects:

Diffusion controlled processes at the Where RF is the release fraction or the ratio between sodium/cover gas interfa~&~-~~ the release rate, R, and the production rate, P, for a nuclide, i.65 From the shape of a curve of measured For Xe-133 and Xe-135, the operation of RFi = (fAi), Brunson derived Ad values for EBR-II.65 the cold trap in the sodium circuit. Expressed in the form of a degassing delay half life rather than a probability, he found Ad equal to 0 Characteristics of the system in question, 7.7 hours. such as mixing and flow conditions. In other measurements, Brunson found a degass- The first case is important for some special experi- ing half-life for EBR-I1 of 2.7 hours, as well as the ments and is not generally associated with LMFBR longer one. Even with relatively small and simple operations. The second case is connected with obser- experimental facilities, such as the fission product vations of some reactors where the concentration of loop described in Reference 64, two half-lives for Xe-133 and Xe-135 in the cover gas could be reduced degassing were found, one between 17 and 32 min- by operating a cold trap in the sodium. It has been utes, the other between 80 and 200 minutes. The shown in this case that the iodine precursors are shorter one can be explained by a defusion controlled adsorbed on the stainless steel surfaces in the cold process at the sodium/cover gas interface. Calcula- trap.64 No explanation is available why the xenon tions of hd from the Xe-135/Xe-135 m ratios resulted formed by the decay of adsorbed iodine remains in even shorter degassing delay half-lives of 4.4 min- within the sodium system. utes at 240°C to 9.6 minutes at 390”C.66

20 The measured delay is actually delay of the total sion product is generally tritium. The average thermal system, including releases from the fuel to the fission yield for tritium in atoms per fission is about sodium, transport within the sodium system, and 0.01 Vo (0.8 x T/fis~ion).~I-~~The major spe- often even the delay within the cover gas system cies released from a thermal reactor is ~~0.~1-72 itself.67 The values given for delay times are therefore always only valid for the system investigated. The An 82-MWt pressurized water reactor cover gas bubble concentration in the sodium will produces 11000 Ci of tritium by ternary fission influence the release behavior very strongly, because and 1380 Ci by boron reactions per ~ear.~()Thefast the exchange of dissolved gases within the bubble is fission yield for tritium from U-235 anlj Pu-239 very intense and the transport of the bubble to the was measured to be 2 x T per fission, which is cover gas is rapid. For example, the EIR-5 reactor con- 2 to 3 times the ternary fission yield for thermal tained large amounts of argon cover gas in the circu- neutrons.71 Other calculations and measurements lating sodium and no delay for the degassing of have shown that tritium yield by fast fission for fission gases was reported. The BOR-60 and BN-350 Pu-239 is 2.25 x that of U-235.70 reactors, on the other hand, did not have such a bub- Two of the potentially more troublesoine paths ble problem, and delay half-lives of about 5.5 and for tritium escape from an LMFBR are the heat 15 hours, respectively, have been published.38, 69 In transfer surfaces of the steam generator syijtem and all other operating reactors investigated, the delay the walls of the containment piping for primary half-lives are in the range of hours, with the exception sodium. Oxide coatings on these surfaces may sig- of DFR, because of its vented fuel elements. nificantly reduce the permeability of hydrogen isotopes by as much as 2 or 3 orders 01‘ magni- Tritium. Tritium generated in the reactor core is of tude.73-76 Directly measured tritium permeation concern because it is extremely mobile and will dif- data for unoxidized and oxide coated metal and fuse in the fuel and control rod cladding and alloys are scarce. In early tritium transport calcula- migrate through the liquid sodium coolant systems. tions, the permeation inhibiting nature of oxide Production of tritium in an LMFBR is greater than coatings was not taken into account when lcomput- that in an LWR at comparable thermal power. The ing permeation losses.77-78 Since, an ctxtensive primary means by which tritium is produced in series of measurements has been made on the per- nuclear power plants are by neutron absorption in meation rate of tritium through unoxidized “clean” boron and through ternary fission. and steam oxidized Croloy (Fe-225 Cr-1 M0).7~-80 The material is a leading candidate for clmstruc- Three boron reactions that produce tritium in tion of LMFBR steam generators. Recent tritium nuclear reactors are as follows: permeation data for oxidized Types 304 and 10 316 SS, from which the sodiu’m containment ves- B (n, CY) T sels and piping will be fabricated, have also been 9 incorporated into newer transport models for trit- B1’(n, T) Be ium.81 Oxide coating to inhibit tritium permeation 10 7 through structural materials is a major feature. B (n, CY)Li (n, na)T. Under normal conditions, all exterior sodium containment surfaces will be oxidized to some degree. The first reaction is the most common in both ther- Therefore, accurate tritium permeation data are mal and fast reactors. The other two are not as favor- needed for both oxidized and unoxidized metals able; Bll(n, T) Be9 has a high neutron threshold and alloys in an LMFBR environment. Protective energy of approximately 9.6 MeV, BIO(n, a)Li7 has oxide barriers are expected to be the second most a small cross section for both fast and thermal neu- effective means of controlling tritium relea $e, cold trons. All three reactions have higher cross sections trapping being the most important. for high energy neutrons than for thermal neutrons, and tritium production by boron reactions will be The efficiency of tritium removal by cold trap- greater in fast than in thermal reactors. Tritium pro- ping strongly depends on hydrogen sources in the duced by irradiation of boron may be the predomi- sodium coolant.82 Because of the predaminant nant source of tritium activity in an LMFBR coprecipitation mechanism in cold traps, large containing boron control rods.70 hydrogen influxes from the steam generator are beneficial; however, these same large influxes also Ternary fission is the fissioning of an atom into one shorten cold trap lifetime. Thus, accurate experi- minor and two major fission products. The minor fis- mental data for hydrogen influx rates are vital. The

21 data would allow cold trap designs to be optimized, increased simultaneously because of excess cover gas and the complex interrelationship between hydro- leakage caused by extensive fuel handling. 84 gen and tritium in the sodium coolant in tritium Q and hydrogen transport models to be characterized The average activity of tritium in the reactor build- completely. At CRBR, production of tritium in the ing air at the 1 m level measured under steady state B4C control rods as a function of rod position was condition was 7.34 f 0.59 x pCi/mL; the calculated using the computer code CRSSA (Con- maximum and minimum activities measured under trol Rod Steady State Analysis). The production steady state were 1.08 f 0.32 x pCi/mL and rate of tritium was calculated to vary from 37 to 4.77 f 0.18 x pCi/mL re~pectively.~~ 92 Ci/d, depending on reactivity control requirements. The average during a fuel cycle from the The tritium that leaks into the EBR-I1 reactor source was calculated to be about 62 Ci/d. building air from the cover gas is mostly unoxidized (T2 and HT), because of the low oxygen concentra- At CRBR, the tritium yield rates from ternary tion in the cover gas and the primary sodium. fission were calculated to be 1.7 x and Greater than 99% of the tritium in air is naturally 2.2 x 10-4 tritium atoms for fission of plutonium converted by isotope exchange and oxidation to and uranium, respectively. The estimated produc- HTO. The percentage of HTO in the reactor build- tion rate from this source is 22 Ci/d. ing air from cover gas leakage varies with the residence time of the tritium in the air before The third significant source of tritium at CRBR sampling. 84 The occupational and general public is from neutron interaction with the lithium con- exposure limits for tritium is based on the chemical taminant in the primary coolant. Two reactions are form HTO, which is %25000 times more hazardous present, Li6(n, a)H3 and Li7(n, n a)H3. The first than T2 or HT.85 The average concentration of trit- of the reactions dominates. The estimated produc- ium activity measured in the EBR-I1 reactor build- tion rate from this source is 5 Ci/d. ing air was approximately four orders of magnitude below the occupational limits for the HTO species. The average production of tritium for CRBR The highest tritium activity measured in EBR-I1 from all sources over one fuel cycle is approxi- reactor building air was approximately three orders mately 89 Ci/d.83 of magnitude below the occupational limits.84

Steady state in the cover gas system at EBR-I1 Tritium activity in EBR-I1 reactor building air is exists when there is no excess cover gas leakage and not currently monitored but can be calculated from the activity of tritium at that time in the cover gas is the cover gas leakage rate and cover gas tritium within la of the norm (2.97 f 0.86 x pCi/ activity, which are both measured routinely. The mL.84 The norm was determined by averaging the tritium activity can also be inferred by monitoring

measured activity of tritium in the cover gas over a other fission products in the reactor building air. ~ 9-month period. The cover gas leak rate was deter- The activities of Xe-135 and Xe-133 are regularly mined by measuring the activity of the gaseous fis- monitored and will exceed occupational limits well sion product Xe-133 in the cover gas relative to the before occupational limits of tritium are exceeded. reactor building air. During normal operating con- For example, the occupational limits for Xe-135 ditions, the cover gas leak is %lo mL/s, the Xe-133 and tritium would be exceeded in 0.41 days and activity in the reactor building is below 1.1 x 5.2 days, respectively, during reactor building iso- pCi/mL, and the Xe-133 activity in the cover lation if the activities of Xe-135 and tritium and gas remains fairly constant. cover gas are at background levels and the cover gas leak rate is 400 times steady state.84 If the Xe-133 Xe-133, but not tritium, is one of the fission gases and Xe-135 activities in the cover gas are decreased released from a ruptured element. The Xe-133 activity an order of magnitude below the equilibrium back- in the cover gas, and therefore the reactor building air, ground levels by decay after shutdown and by selec- increases rapidly during the release, whereas tritium tive removal during purification of the cover gas, activity in the reactor building air and cover gas tritium activity in the reactor building air cannot be remain steady. In some instances, the activities of inferred from the Xe-133 or Xe-135 activity in the Xe-133 and tritium in the reactor building air have reactor building air.

22 Radioactive Waste Management duced by the reaction of water/alcohol with the radioactive sodium that adheres to items. The Liquid Waste System cleaning is done in the sodium component maintenance shop. Design Objectives. An LMFBR liquid radiological waste system is designed to process contaminated Tritium is produced in EBR-I1 and Nwentually liquids prior to reuse or release into the environ- released in the liquid ~tream.~~-~~Ternary fission ment. As with an LWR, design objectives of an is the major source of the tritium, which is released LMFBR are to purify and reuse waste liquids where to the primary sodium coolant. Althoggh more possible and to minimize the total activity in liquid tritium is generated in the boron carbide control effluents in the total volume of concentrates that rods and is present in materials studied in experi- will require drum or other storage. The basic mental subassemblies, it is quantitatively retained approach is to process liquid rad waste so that vir- in the boron carbide at the temperatures reached in tually all radioactive materials are contained in EBR-I1 and is not released to the coolant. Approxi- solid material, to load all the solid radioactive mately 90-95'70 of the tritium produced in ternary material into containers that meet Department of fission is released from the fuel through the SS Transportation regulations, and to transfer the con- cladding to the primary sodium coolant.t16 tainers to a licensed contractor for processing or disposal. Under normal operating conditions, any The level of tritium in the steam-turbine conden- radioactive waste released will be as low as is rea- sate averages about 10 pCi/cm3. The makeup rate sonably achievable and less than the limits sets by for the steam system is 38 m3 of water/day; there- 10 CFR 20. fore, the tritium release through the steam system is about 380 pCi/d. At a plant factor of 70%, this The source of low level waste at LMFBRs are totals about 0.1 Ci/y. drains for floors, equipment, laboratories, personal decontamination showers, and maintenance The Prototype Fast Reactor (PFR), the Fast shop. Sources of intermediate level waste are fis- Reactor Fuel Processing Building, and associated sion products, fuel, and corrosion products having laboratories and support facilities at the Dounreay plated out or deposited on components cleaned in Nuclear Power Development Establishment pro- decontamination facilities. duce low to high level liquid radioactive waste. The high level liquid radioactive waste contains fission Releases from LMFBRs. The liquid radioactive products from the fast reactor fuel reprocessing waste from PHENIX consists of wash water from building. They are transferred to shielded tanks in the washing of fuel elements prior to shipment to the reprocessing plant for sampling and analysis the reprocessing plant, and the washing of compo- and then dispatched to underground storage tanks nents withdrawn from the reactor core prior to that are continuously cooled to remove the heat repair or storage. These liquid wastes are trans- generated by the decaying fission products. The ported by special tank trucks to the waste treatment activity of these liquid radioactive wastes is reduced station at Marcoule. As of April 1980, shipments by a factor of about 500 during the first five years to the Marcoule plant represented some 2700 m3, of storage. All remaining aqueous wastes are col- corresponding to an activity of 580 Ci. Liquid lected in storage tanks in the solvent extraction waste from the PHENIX plant represents plant where they are monitored, sampled, rind ana- '~1500L/MWyt or 0.18 mCi/MWyt of liquid lyzed. They are then piped to vessels for treatment waste production. by a flocculation process. A slurry containing the radioactive material is produced and separated No radioactive liquid waste is produced by oper- from the water, and is transferred to settling tanks ation of the EBR-I1 reactor or within the contain- where it is stored for eventual recycling. The liquid ment building except for controlled liter batch residue is piped and held in tanks, where its activity quantities of water/alcohol used for component is monitored to ensure that it is within the liimits set decontamination.86 Therefore, no liquid waste sys- by the Secretary of State for Scotland before being tem has been installed within the containment discharged to sea through an underground pipe- building. Components and equipment to be line. The liquid releases are limited to less than repaired or discarded are cleaned to remove or react 6000 Ci of alpha and beta emitters in any three con- the attending sodium. Radioactive liquid is pro- secutive months, and in which no more than 600 Ci

23 can be strontium-90 or 60 Ci transuranics. Studies treatment systems, and the extent to which these show that release of these liquid aqueous wastes systems are used. EBR-11, for example, generates become associated with fine suspended nonorganic liquid waste not on the basis of power produced, particulate material, which adheres to the nets of but on the number of subassemblies removed and fishermen. The fishermen dry their nets on the washed. The data for the Phenix reactor includes beach for a period of 36 h, which allows beta emit- the period during which repair work and changes ters, Ce/Pr-144 and Run/Rh-106 to become fixed were made to the damaged intermediate heat to the nets. The maximum calculated dose to the exchanger, which produced large quantities of liq- fisherman’s hands from these fixed beta emitters is uid waste during a reactor shutdown. Nevertheless, 5 rem.88 the overall environmental release from liquid waste per unit of energy generated is considered to be a Current operating procedures at LMFBRs for reasonable index for comparing an LMFBR with a liquid rad waste systems are analogous to those commercial LWR. The data in Table 11 demon- being used at LWRs. The basic operations in the strate that present LMFBRs produce orders of systems are evaporation and ion exchange. These, magnitude less liquid waste quantities per mega- and ancillary collection, storage, and fluid trans- watt year thermal power production, and one to port operations, separate the activity from the liq- several orders of magnitude less activity in their liquid waste. The detailed procedures that are part of uid waste, as a function of power produced. the LWR technology and found in Reference Safety Analysis Reports for standardized LWR plants, Gaseous Waste Systems and in-plant procedures for operating LWRs, are adequate also for LMFBRs. Types of components Design objectives. LMFBR design objectives for a in LMFBR liquid waste systems, including capac- gaseous waste system at, like those of LWRs, are to ity, flow rate, size, weight, design code, and seismic keep the level of radioactive material in the plant classification are similar to those normally used in effluence to the environment as low as reasonably LWR plants. Process and radiation instrumenta- achievable. Extensive effort has been made in tion used to measure the flow, pressure, tempera- developing system designs that have resulted in ture, liquid level, radiation level, and pH of the minimizing or eliminating gaseous release of radio- liquid rad waste stream at LMFBRs were found to active material to the environment during normal be the same as those used at LWRS.~~Table 11 plant operations. LMFBR plant design objectives compares radioactive liquid releases from LMFBRs in the United States include conformance with the with those from the average US commercial LWR. requirements of 10 CFR 20. The radioactive gas- Caution should always be taken against making eous releases at LMFBRs compare favorably with simplistic comparisons of radioactive releases with those at commercial nuclear power plants. Radio- the energy generated, because of the many factors active airborne effluence from LMFBRs consist of that affect the amount of radioactive materials ,noble gas fission products (which include isotopes released, such as the condition of the fuel, primary of xenon and krypton and their daughter prod- system integrity, effluent and radioactive waste ucts), activation products entrained in the argon

Table 11. Comparison of LMFBR liquid radioactive waste production with U.S. PWRs

Radioactive Waste Produced

Volume of Liquid Activity of Liquid Reactor [L/MWY(t)l [Ci/M Wy(t)]

Phenix 1,372.0 0.29

FFTF 71.5 1.58E-5

EBR-I1 3,811.0 3.07E-3

U. S . P W R (average) 54,000.0 2.795

24 cover gas, and tritium. Tritium is the most signifi- product concentration in the reactor building. The cant radioactive gaseous effluent at an LMFBR. purge-discharge rate (a maximum of 0.08 m3/min) Tritium is generated during operation primarily by was determined by identifying the nature and con- ternary fission and by activation of B4C control centration in the containment, so that discharges rods. A small amount of tritium is generated by did not exceed acceptable limits. The purge exhaust neutron activation of lithium and boron, which are during this time bypassed the HEPA filters and impurities in both the sodium coolant and the core went directly into the stack. and blanket fuel. Tritium behaves much like hydrogen and, at the elevated temperatures found in The concentration of radionuclides in the con- LMFBRs, diffuses in stainless steel and other reac- tainment building increased in proportion to the tor containment material. It is, therefore, of partic- activity of the argon cover gas. Under off-normal ular importance to closely monitor the tritium in conditions of cover gas activity, samples from the the coolant system and to discern the relationship containment building were taken and analyzed. If between the tritium concentration and the various the activity level was within predetermined accept- reactor operating parameters. Tritium transport able limits, the containment building atmosphere through an LMFBR, and the performance of trit- was purged directly to the outside atmosphere ium control methods, must be carefully monitored through a 3 m3/s centrifugal blower. (This was in to minimize tritium release to the environment. addition to that processed through the HEP4 filters.)

Releases of Gaseous Radioactive Waste. The major In 1977, anticipation that breached elements in portion of EBR-I1 gaseous waste releases prior to the run beyond-cladding-breach (RBCB) program 1970 originated from the argon cover gas system. would release substantial quantities of fission prod- The reactor core is immersed in a primary containment vessel that contains approximately 325 m3 of ucts to the primary system, especially fission gases, the cover gas cleanup system (CGCS) was installed. molten sodium. Immediately above the surface of The system processes primary cover gas to remove the sodium is an 18-m3 plenum region filled with fission gases in a cryogenic column. Effectiveness argon. Because of a positive pressure differential, of the system, and substantial reductions in the gas any leakage gas is combined with the atmosphere in leakage rate through the primary tank col'er to the the containment building. Approximately 2.5 m3/s building containment, sharply reduced the release of the building atmosphere is withdrawn through of fission gases to the environment via the waste gas the shield cooling system and through HEPA fil- stack, as shown in Table 12. This reduction applies ters, and combined with another 0.75 m3/s that is to both the RBCB and RTCB programs, which are withdrawn through the thimble-cooling system and being conducted simultaneously. also passed through HEPA filters. The combined flow of 3.2 m3/s is then passed through a radiation monitor and through a blower to the 60-m high The CGCS became operational in June 1977. stack. Approximately 2.1 L/min of argon cover Releases from EBR-I1 in 1977 prior to CGCS oper- gas from the cover glass plenum is discharged ation totaled 41 1 curies. Although the RBCB pro- through monitoring devices into the radioactive gram results in greater fission gas releases to the gaseous waste disposal system downstream of the cover gas, releases after the startup of the CGCS HEPA filters. Table 12 presents annual releases of totaled only 50 curies. Prior to June 1977, Xe-135 radioactive gases from the reactor containment and Xe-133 comprised the large majority of the building for the years 1979 through 1981. radionuclides releases to the environment via the EBR stack. Since that time, approximatt:ly three From 1970 to 1976, there was an expansion of the fourths of the radioactive gaseous effluent has con- testing of fuel elements in the run-to-cladding- sisted of Xe-135 and Xe-133, and one foiirth has breach (RTCB) program. Such operation produced consisted of Kr-85, as shown in Table 12. breached fuel elements, and the fission product and noble gas concentrations increased. In order to Tritium is produced in EBR-I1 by ternary fissions in establish the identity of the subassembly containing the core, and by neutron bombardment of the B4C the failed elements, continued reactor operation portion of the control rods. Tritium production from was sometimes necessary, thereby increasing fission B4C can be ignored when evaluating tritium transport product concentrations in the argon cover gas and in EBR-11, since essentially 100% of the tritium pro- the containment building. It then became necessary duced from B4C is retained in the control rods at to purge the cover gas system to reduce the fission ,\ EBR-I1 temperature^.^^ The ternary fast fission yield

25 I Table 12. EBR-II radioactive airborne effluents n

Total 1981 Nuclide T1/2 Ci

Xe-133 5.3 d 39 Xe-135 9.2 h 13 Kr-85 10.7 y 15

I 77

1979-100 Ci 1980-130 Ci 1981- 77 Ci

in EBR-I1 for U-235 is reported to be 2 x T/ inert cells surrounding the piping, (b) permeate fission, or 1 atom of tritium per 5000 fissions. The through the intermediate heat exchanger (IHX) thermal fission rate is 2.7 x lgl fissions/MWd, so tubes into the secondary circuit, (c) escape by leak- the production rate of tritium at EBR-I1 at full power age from the primary system cover gas, (d) copre- (62.5 MWt) is 3.4 x atoms of tritium/day, or cipitate with hydrogen in the primary cold traps, 1.65 Ci/d.77,84 and (e) decay to He-3. Tritium that reaches the secondary circuit through the IHX may (a) permeate As shown in Table 13,90-95% of the tritium pro- through the secondary sodium containment walls duced by ternary fission in EBR-I1 is transferred to into the surrounding air cells, (b) permeate the primary sodium and is dispersed throughout through the walls of a steam generator tubes into the primary circuit by bulk transport of the cool- the water/steam system, (c) escape by leakage from ant. From the primary circuit, tritium may (a) per- the secondary system cover gas, (d) coprecipitate meate through the SS sodium containment into the with hydrogen in the secondary cold traps, and

Table 13. Summary of EBR-II dataa of tritium distribution

Tritium Typical Sample Activity Location (VO) (Ci)

Fuel 5-10 -

Primary sodium 4 15.5

Primary coldtrap 83-88 -

Areas outside primary sodium system:

Primary argon 0.0015 Secondary sodium 0.083 Secondary argon 0.00018 Steam system 0.0005 8

a. Data taken between 1972 and 1974.

26 (e) decay to He-3. Approximately 83438% of the Previous to the experiment, because data was produced tritium is removed by the primary cold lacking, the Sievert’s constant for tritium and L.$ trap .84 sodium was assumed to be equal to that for hydrogen and sodium.77 The Sievert’s constant for trit- -. The rate of permeation through the sodium con- ium in sodium has been estimated on the basis of tainment walls is a function of wall thickness, the decomposition pressure in the analogous LiH, surface area exposed to tritium, the tritium concen- LID, and LiT systems, and the associatcd isotope tration in the sodium coolant, the Sievert’s law con- effects .92 stant for tritium in sodium, and the permeability of tritium through the reactor construction materi- The rate of tritium permeation through the A transport model has been used to predict trit- walls of the IHX is also a function of these same ium and hydrogen distribution in CRBR. Rates of variables; however, the driving force for the process tritium loss from reactor containment materials is caused by the difference in tritium concentration have also been calculated.82 The predict1:d tritium in the primary and secondary sodium circuits. Two and hydrogen profiles and release rates are strongly of the potentially more troublesome paths for trit- affected by two factors: (a) the permeation- ium escape from an LMFBR are the heat transfer retarding nature of oxide coatings, particularly on surfaces of the steam generator system and the stream generator tubes and sodium coolant piping, walls of the containment piping for the primary and (b) the rate of hydrogen ingress to the second- sodium. It is well known that the outside coating on ary coolant circuit caused by steam gene rator cor- these surfaces may significantly reduce the permea- rosion. In regard to the first factor, a reference case bility of hydrogen isotopes by as much as two or was established in which all structural and heat three orders of magnit~de.~~-~~,~~However, transfer surfaces were assumed to be oxide free. directly measured tritium permeation data for Maximum tritium and hydrogen permeai ion rates unoxidized and oxidized coated metals and alloys through the construction materials were obtained, are scarce. as well as upper limit values for tritium release rates from reactor subsystems. In regard to the second An experiment was carried out at EBR-I1 to pro- factor, the effectiveness of cold traps directly vide data for which the tritium permeation rate depends on the hydrogen ingress rate from the across the IHX was determined. The intermediate steam generator. Since tritium is most el’fectively heat transport system cold trap of EBR-I1 was removed from the coolant by coprecipitate with bypassed from May 20 to June 8, 1977, and trit- hydrogen in the cold traps, large hydrogen influxes ium concentration increases in the primary heat from the steam generator should produce high cold transport systems (PHTS) and the intermediate trapping efficiency. However, the useful lifetime of heat transport system (IHTS) sodium were sodium cold traps is greatly reduced by large hydro- recorded with in-sodium tritium meters (ISTM) gen influxes. Cold trap designs must be optimized, that had been recently installed in the rea~tor.~O-~l and anticipated tritium and hydrogen burden in The tritium flux across the tube wall was calculated both primary and secondary cold traps must be from the physical properties of the construction determined. material, such as area, thickness, and tritium permeability in Type 304 SS. At the Phenix reactor, the gaseous waste dis- The tritium and hydrogen distributions were pre- charge from 1974 to 1981 totaled only 970 ci, for dicted for the 9 days during which data were an average value of about 0.5 Ci/day. Nearly all of recorded. The calculation represented a transient this activity was owing to Xe-135. In 1953, Phenix case, since the secondary cold trap was off-line dur- began testing a cryogenic cleanup system for xenon ing the entire period. Figure 1 compares the results radioisotopes. Currently, the Phenix off-gar; system of this simulation with the experimental data for includes only a delay line. The volume of the cover tritium and hydrogen levels in the PHTS and IHTS gas in the Phenix reactor is only about 40 m3, with sodium of EBR-11. The solid lines represent the a radionuclide concentration of approximately profiles predicted with the transport model; the 5 Ci/m3, most of which is from xenon and lcrypton symbols represent the actual operating data. radioisotopes. The only activation products Hydrogen concentration data of the PHTS were detected were Ne-23, 1.5 x 10-1 mCi/niL and not available. Agreement between the calculated Ar-41, at 5 x mCi/mL. At this time the and experimental values is good for the time inter- Phenix reactor does not have a cover gas cleanup val studied. system.93

27 Time (days)

Figure 1. Tritium and hydrogen concentration in EBR-I1 sodium coolant systems.

At the Prototype Fast Reactor (PFR), the most During the cycle-1 operation at FFTF, four noble significant radioactive gaseous effluent is Ne-23. gas releases occurred that were above background. Total radioactive gaseous effluents emitted from These include the following: PFR are equivalent to 4-5 derived air concentration (DAC) hours per year.94 Q Approximately 116 mCi of noble gases were released via the lower reactor service build- KNK-I1 releases are, on average, approximately ing (RSB) exhaust over a 21-hour period on 10 Ci/yr. Nearly 400 Ci, mostly Xe-133, were May 22, 1982. The source of this release was released in April 1979 during the first fuel leakage a leaking instrument tubing connection. incident. At full power and operating with no defective fuel pins, KNK-I1 releases were 180 mCi Approximately 1.1 Ci of noble gases were Ar-41/m3, 2-3 mCi Zn-135/m3, 1 mCi Zn-133/m3, released via the combined containment and 0.5 mCi Kr-87, Kr-88, Kr-85m/m3, all believed exhaust over the period of May 15-28, 1982. from the presence of tramp uranium, for the emission The source of this release was the end con- rates are not linear but an exponential function of tainment cell atmosphere processing system reactor power. 95 (CAPS) flood exhaust.

28 About 60 mCi of noble gases were released The system should be designed to handle, pack- from a cover gas leak in a gas tank sample age, and store three sources of wastc concen- trap connection. The release occurred inter- trated liquids, compactible solids, and noncom- mittently over a 15-hour period on pactible solids. It should be designed to perform its September 28, 1982. The total release time function with radiation hazards to plant personnel was 5 hours. as low as reasonably achievable through the use of proper layout and arrangement, coupled with ade- Approximately 12 mCi of noble gases were quate shielding. Equipment should be selected, released when an O-ring was inadvertently arranged, and shielded to permit operation, inspec- omitted from a reactor cover gas grab sam- tion, and maintenance, with minimal exposure to ple cylinder connection. This release personnel. All operations should be monitored to occurred over a 3-hour period on ensure that radiation concentrations in accessible October 29, 1982. areas are below the exposure 1imii.s set by 10 CFR 20. The tritium concentration in the primary cover gas and interim decay storage cover gas at FFTF is The design of solid rad waste systems are based in the range of 10-1 to Ci/mL. Samples of on widespread experience in the nuclear industry, the secondary cover gas impurities at FFTF have spanning more than two decades. Reference 97 tritium concentrations in the range of to contains a review of solid rad waste practices at Ci/mL.96 Radioactive gaseous effluents at nuclear power plants. In particular, the imorpora- FFTF since Cycle 1 have averaged %O. 117 Ci/ tion of concentrated liquid rad waste in cement is MWy thermal. widely and routinely used in power reactor plants. The packaging of solid material for disposal by During the years 1963 to 1979, EBR-I1 has emitted burial is also standard practice. an average of 19.6 Ci of noble gases per/MWy, and , 39.18 Ci tritium/MWy. Since the start of the run-to- The operating procedures for handling the five 9 cladding-breach (RTCB) program in 1973, total radio- types of input streams in a solid rad waste system at active gaseous releases have increased to about an LMFBR are as follows: 120 Ci/MWy. In anticipation that the breached elements in the run-beyond-cladding-breach (RBCB) Concentrated liquids, which include evap- program would release significantly higher quantities orator bottoms, spent resins, sodium con- of radioactive gaseous effluents to the primary sys- taminated ethyl alcohol, and tritiated tem, the cover gas system, and the environment, a water, are solidified. A fixed amount of cover gas cleanup system was installed in 1979. Subse- liquid rad waste can be metered into quent to installation of the cleanup system, radioac- 55-gallon drums that have been preloaded tive gaseous effluents from EBR-I1 have dropped with cement. The drum can be capped, considerably, to about 18.6 Ci/MWy. thoroughly mixed by tumbling, decontaminated by rinsing, monitored, stared tem- Radioactive gaseous effluents at commercial porarily if need be, and transferred to a nuclear plants from the years 1960 through 1976, licensed contractor for disposal. have averaged 35.8 Ci/MWy. This is significantly higher than the radioactive gaseous effluents emitted Compactible solids such as rag!;, paper, from LMFBRs. and rubber seals, can be collected at various points throughout the plant and trans- - Solid Waste System ferred to a solid rad waste system, as is done at LWR plants. Design Objectives. Solid rad waste systems should be designed to process and package solid wastes Noncompactible solids include used sup- that are to be shipped off site for disposal. Packag- port tools, contaminated filters, metal ing must be such that the surface dose rate will be in from cutting operations such as an inter- compliance with the Department of Transportation mediate heat exchanger tube bundle, regulations. valves, and vapor traps. The low activity

29 noncompactible solids can be placed in 55- shielding in a metal container underground in an gallon drums, capped, decontaminated, interim storage facility at Argonne National monitored, and placed in temporary stor- Laboratory-West (ANL-W), designated as the age. Spent filter cartridges from the liquid radioactive scrap and waste facility. Storage of the rad waste system can be remotely placed in cold traps is awaiting development of a treatmect . concrete line 55-gallon drums and placed process. in storage prior to shipment in the same manner as other noncompactible solids. Most of the solid radioactive waste is generated at Radioactive sodium will be present as a the hot fuels examination facility (HFEF). This waste result of fuel handling. This metallic results from the disassembly, assembly, and inspec- sodium can be transferred to the rad waste tion activities of EBR-I1 subassemblies, discarding of system in 55-gallon drums. Since no waste used experimental hardware, preparation of reactor disposal site will accept sodium, it is blanket subassemblies for storage, and inspection placed in temporary storage on site or activities of LMFBR related tests that are performed processed into a nonmetalic form, such as by other organizations. Solid waste is also produced a carbonate or carbide, and stored on site. when the evaporator bottoms are solidified. The solid wastes are of the following categories: low-level non- The primary and intermediate cold traps are transuranic, intermediate nontransuranic (up to the sources of sodium bearing waste. Pri- 10,OOO R/h), low-level transuranic, intermediate level mary cold traps should be designed for the transuranic (up to lo00 R/h), low and intermediate life of a plant and not expected to be level transuranic, and nontransuranic waste with removed. If replacement of a cold trap sodium, and bulk sodium at the intermediate level. becomes necessary, it can be drained of bulk sodium and stored in a designated location. The chemistry laboratories that support EBR-I1 Intermediate cold traps are not designed for produce solid waste from maintenance and opera- the life of the plant, will require removal, tions. These wastes comprise low and intermediate and storage must be provided. levels but the volume is low and there is no elemental sodium. Solid Waste System Experience at LMFBRs. At EBR-11, a small quantity of solid waste is produced. A major portion (by volume but not by radioactivity) Radioactive solid waste is handled by several differ- of this solid radioactive waste is the accumulation ent methods, depending on the radiation levels and of wipe rags, plastic containers, shoe covers, and the content of the waste. Low-level nontransuranic other industrial solids associated with working compactible waste is collected in plastic bags and with radioactive materials during maintenance. hand carried to and placed in specific strategically Reactor components such as thermocouples, nuts located dumpsters. The dumpsters are then trucked to and bolts, and other hardware are disposed of as the INEL Radioactive Waste Management Complex solid radioactive waste. Radiation from these com- (RWMC) operated by EG&G, Idaho, where the waste ponents is generally low level (<10 mR/h). is compacted prior to storage. EG&G Idaho has reported compaction greater than 10: 1. Another source of solid waste is elemental sodium in bulk form, which is produced in small quantities during maintenance. This sodium is col- Prior to December 1977, low-level nontransuranic, lected in 3.8-L cans, which are placed in metal noncompatible waste was placed in cardboard boxes; drums and covered with sand. Because of low radi- the boxes were placed in dumpsters and sent to the ation levels, these drums are stored in a covered RWMC. Bulky or heavy items were boxed in wooden trailer until the sodium can be treated. Approxi- boxes. These boxes were then sent to the RWMC for mately 500 L of elemental sodium are presently in storage. Since December 1977, this low-level waste

~ cold traps used in the reactor coolant system to con- has been collected in 208-L steel drums. The drums trol oxygen. The traps are removed from the system are held in a cupboard van until the van is full and when their efficiency has dropped below preset lev- then sent to the RWMC. Items too large for the drum els, and they are stored in a radioactive storage are placed in standard-sized plastic lined wooden building if the levels are low, less than 100 mR/h. boxes painted with fire retarded paint. The boxes are When the radioactivity is high, they are stored with available in sizes up to 1.2 x 1.2 x 2.4 m.

30 Low-level transuranic waste is placed in Depart- Attempts have been made to reprocess the sodium ment of Transportation (DOT) specification 17C from the Rhapsodie reactor to remove the fission drums equipped with a 2.3-mm rigid poly liner, or products, but with very little success.93 in fiberglass DOT specification 19A boxes. The filled containers are sent to the RWMC for storage. Whether sodium waste is disposed of in a landfill However, no transuranic waste has yet been gener- or in a permanent depository, it will be necessary to ated as a result of EBR-I1 operations. convert it to a nonmetallic, stable compound to minimize its subsequent reaction and transport in Intermediate level waste with radiation levels up to the environment. Various types of glasses contain- 10,OOO R/h is generated within HFEF as a result of ing silica and sodium monoxide may be suitable as the disassembly of EBR-I1 driver subassemblies prior such stable compound. For example, the composi- to inspection or shipment of fuel to the reprocessing tion of ordinary window glass is 17% Na20, 6V0 plant. Intermediate waste from HFEF is remotely CaO, 1% Al2O3, the remainder being Si02. The packaged in 1.8-m long, 295-mm ID carbon steel volume of this glass made from a given mass of cans. These cans are then inserted into a stainless steel elemental sodium is approximately 3 times the outer can and either seal-welded or gasket-closed. The original volume of the sodium. cans that do not contain elemental sodium are sent to the RWMC for interim storage. Cans containing ele- Ordinary window glass is not ideal for disposal mental sodium are sent to ANGW, RSWF, for stor- of radioactive sodium from the standpoint of age. No intermediate level transuranic waste is leaching of fission products by water. ]However, generated by EBR-11. other glass compositions have been developed as candidate materials for encapsulation of high-level A total of 9.7 m3 of solid radioactive waste, with waste from fuel reprocessing.98 a total activity of 19.8 mCi, was transformed from FFTF to the 300 area at Hanford for processing. These glasses typically contain both silica and The waste originated from the following activities: sodium monoxide in various Si:Na ratios. 1 t may be efficient to use the sodium monoxide from sodium Interim examination and maintenance waste disposal as one of the feed materials for pro- (IM) cell decontamination and mainte- duction of special purpose glasses in the disposal of nance radioactive sodium.

Maintenance and storage facility decon- The processes that would be suitable for convert- tamination and maintenance ing radioactive sodium to an acceptable compound for waste disposal must meet certain performance Sodium removal system maintenance criteria related to the presence of radioactive contaminants and transuranics. These criteria include Heat transport system south sampling and (a) low cost, (b) simplicity, to allow remote opera- maintenance tion, (c) safety, to prevent uncontrolled Ieaction, (d) suitability for incorporation into a disposable Miscellaneous maintenance in the plant form, and (e) minimum release of radioaci ivity.

Radioactive liquid waste sampling and Disposal of sodium and other alkali metals have load out station. been accomplished by a variety of methods, depending on the requirements of each situation. At the Phenix reactor, low level waste is handled essentially like as it is at EBR-11. Intermediate waste The most frequently used methods have been is transported to the Marcoule center and stock- (a) reaction with alcohol, (b) reaction with steam or piled in pits. The total quantity of these trans- water vapor, (c) reaction with concentrated caustic ported intermediate wastes to date amounts to solutions, and (d) burning. The alcohol reaction about 60 tons, and corresponds to an activity of processes have been used extensively for removing about 8 million Ci. The irradiated fuel is trans- sodium from reactor components, but the quantities ported to the chemical reprocessing plant at processed have been small to moderate, in the 0.1- to La Haque or at Marcoule. Approximately 50-kg masse^.^^-^^ Data were obtained on reaction 100 metric tons of solid sodium is storaged on site. rates and solubility of sodium in alcohol. The initial

31 reaction rate increases exponentially with tempera- range of experience has been gained in actual ture; at 335 K the rate is 0.064-k sodium/m2-s on a sodium-cleaning operation at reactor facilities horizontal sodium surface. 99-108However, the rate using this diluted water method. 'The disadvantages decreases precipitously to essentially zero when the of the method include the liquid waste (NaOH sodium concentration in alcohol reaches 5 wt%. solution in water) that must be dried for disposal, and the evolution of hydrogen and carry-over of Some factors that could limit the use of the alco- moisture in the gas stream. This hydrogen and hol reaction process are as follows. The sodium moisture must be treated appropriately to remove content of the alcohol must be kept low, thereby fission products before venting to the environment. requiring continuous regeneration of alcohol by distillation during the reaction. The insoluble pre- Third is reaction with a concentrated caustic cipitates from the distillation must be further proc- solution and is related to the process developed for essed, and difficulties arise in reducing the product direct reaction of sodium with water.lol Liquid

c I to'a solid because of the presence of a variety of sodium (or NaK) is dispersed in a pool of concen- organic compounds. Disposal of liquid waste con- trated (14 M) NaOH while water is periodically taining radioactive isotopes by landfill burial is not injected to control the caustic concentrations. This acceptable; therefore, disposal of the waste from caustic solution process was used to dispose of the the alcohol process presents formidable problems. NaK from the decommissioned EBR-I reactor.lo3- Other factors include the flammability of the alco- lo4 The NaK was injected into the caustic with a hol and the necessity for processing the sodium in spray nozzle designed to atomize the NaK in a batches. stream of nitrogen. The disposal operation was accomplished smoothly. However, converting the In the early days of sodium technology, some process for operation in a shielded, alpha contain- quantities of sodium were reacted directly with ment facility would present difficulties relating to water by dumping the sodium into a large outdoor moisture, hydrogen handling, and liquid waste water pool. This method, although effective and disposal. rapid, is obviously unsatisfactory for radioactive sodium. An improvement in the dumping method The sodium burning process has been used exten- of the sodium in water was developed for disposal sively, both in open areas where the smoke was of large quantities of sodium under water. lol Liq- released to the environment and in closed facilities uid sodium was sprayed into water at a sufficiently where the smoke was removed by water scrub- high Reynolds number (greater than 45,000) to bing.101*105,106 Although the burning process is ensure disposal of the sodium. The depth of injec- effective in consuming metallic sodium, the sodium tion was approximately 3 m, to ensure complete monoxide smoke necessitates cleaning operations reaction before the metallic sodium could reach the after the burning is complete. Large volumes of gas surface. Adaptation of this process for use in a must be treated to remove fission products and liq- shielded, alpha containment facility, may be diffi- uid waste from the scrubber. Products from the cult because of the need to control gas and water wash down operations must be treated to form solid flows and to handle the hydrogen and water waste. The burning chamber also requires periodic evolution. cleaning to avoid accumulation of prohibitively high activity levels; this cleaning introduces an The sodium water reaction can be controlled by additional difficulty into the method. diluting the water and thus controlling the rate of access of water to the sodium. Two methods of Dissolution of sodium in a heavy metal such as lead water dilution have been used successfully, namely, or mercury, used in the USSR for cleaning reactor the use of steam and the use of water vapor and components. A process using this concept was pat- nitrogen. lo2 These processes have been widely ented in the United States as a method for dis osing used throughout the world, primarily for removal of large quantities of contaminated sodium. lJ7 The of residual sodium from components to be reused process is designed to divide the large heat of reaction in the reactors. Methods of this type include (a) the into three approximately equal energy steps: (a) dis- use of moist argon or nitrogen at EBR-11, (b) the solution of sodium in lead, (b) contacting the lead- use of an atomized water spray in France, (c) steam sodium alloy with molten NaOH and sparging with cleaning in the U.S., Germany, the UK, and the oxygen to drive the sodium (as Na20) into solution in USSR, and (d) the use of water vapor in nitrogen as the NaOH, and (c) injecting steam into the molten practices in the U.S., UK, and Germany. A wide NaOH to convert the dissolved Na20 into NaOH. n

32 The molten NaOH is then drawn off periodically for power supplies, microprocessor and accessories, disposal as solid waste. An attractive feature of this and indication and local alarm indicator lights. i process is that there are theoretically no effluent gases Each monitor is capable of transmitting riidioactiv- and no aqueous waste to be processed. However, the ity level and alarm status information for display process requires high temperatures (> 700 K) and and logging in the control room with redundant complicated liquid flow manipulations. display and logging equipment, and are designed in compliance with IEEE 279-1971. Peric'dic sam- At Argonne National Laboratory West, two pling of primary sodium, secondary sodium, ex- methods have been developed and tested on a small vessel sodium, and cover gases is conducted to alert scale for converting sodium waste to inert com- the operator of any abnormal conditions that may pounds suitable for disposal. The first method is be developing. Both local and remote liquid sam- direct oxidation of the sodium after dispersal in a ples are taken. Isotopic analysis of process samples silica matrix. The sodium is mixed with silica and at LMFBRs in the United States is performed by oxidized in a rotary drum reactor. The product is automated spectrum analysis systems that satisfy suitable for making glass when other stabilizing the reporting requirements of Regulatory compounds are added. The second method is the Guide 1.21. reaction of elemental sodium with molten NaOH at 450°C and subsequent injection of steam into the LMFBR Experience. The fission products released melt to convert the reaction products (Na20 and from a fuel element with breached cladding in NaH) to additional NaOH. The reaction is smooth EBR-I1 have been monitored in several ways. and easily controlled, with a low probability of run- Before 1972, the fuel element failures were caused away reaction. The end product is molten NaOH, by defective closure wells and the like. Although which can be cast into drums for further treatment xenon tags were included in experimental elements or disposal. The advantages of these two methods after 1969, none of the early failed elements had over more conventional aqueous processes are the them. The early period was characterized by elimination of aqueous wastes and the elimination lengthy searches for sources of fission gas release. or minimization of gaseous effluents. lo* Another methodology being investigated at EBR-I1 for the From 1973, intentional endurances, or run-to- disposal of radioactive sodium waste is based on cladding-breach (RTCB) tests produced an increas- converting it to a carbonate or carbide form.lo9 ing number of failures. Generally, however, these failures were well separated in time and did release Process and Effluent Radiological Monitoring xenon tags. Experience was gained in correcting the Systems tag compositions or changes caused by in-reactor exposure, and identification was quite rapid. Design Objectives. Process radiation monitors are placed at LMFBRs to evaluate plant equipment From about 1977, when reference FFrF fuels performance and to measure, indicate, and record tests approached radiation exposure goals, RTCB the radioactive concentration in plant process and tests of advanced fuel element design we re begun effluent streams during normal operation and and run-beyond-cladding breach (RBCB) testing anticipated operational occurrences. Radiation was started, and incidents of failure sharply monitoring of process systems provides early warn- increased. During this period, an average of 10-15 ing of equipment malfunctions indicative of poten- failures occurred per year. The period wa:, charac- tial radiological hazards. Monitoring of liquid and terized by an increasing number of occasiclns when gaseous effluents under normal operating condi- several failures occurred in the reactor at 1 he same tions are done in accordance with NRC Regulatory time. Guide 1.21. The number, sensitivities, ranges, and locations of the radiation detectors can be deter- The ways in which breached elements have been mined by requirements of the specific monitored identified and removed from EBR-I1 from 1967 to process during normal and postulated abnormal April 1979 are described in References 11 C1 to 12 1. conditions. All radiation monitors should be designed so that saturation of detectors during The main cover gas activity monitor now in use severe accident conditions will not cause errone- at EBR-I1 is the germanium lithium argon ?.canning ously low readings. Continuous radiation monitors system, or GLASS, which was developed from pre- presently found at LMFBRs are equipped with vious ~ystems.l~~-l~~Argon is extracted from a

33 cover gas, aged for about four minutes to allow the which arises from decay iodine in the sodium bond Ne-23 activity to decay away, and monitored in a that escapcs before the stored fission gas. 25-mL chamber above a coaxial Ge(Li) gamma detector. Signals are fed into a dedicated computer Any delayed neutrons (DNs) released from a used as a 1024-channel analyzer. The activities of breached element are detected in a bypass stream of Kr-85m, Kr-88, Xe-133, Xe-135, Xe-l35m, and Xe- sodium by the fuel element rupture detector (FERD), 138 are routinely monitored by the GLASS. which was upgraded in March 1978.124-126The new FERD system is improved slightly in counting effi- During normal full power operation, there is a ciency and, more importantly, exhibits a more stable little cover gas activity, because of the fissioning of and reproducible behavior. Detection is accomplished 2-5 mg of “tramp” uranium. This background has by BF3 counters embedded in a moderator stack that remained constant and is thought to be caused by surrounds a sample duct containing primary coolant some fixed contamination in the core that resulted aged about 20 seconds. A FERD loop flow reduction from use of reprocessed driver fuel elements technique has been ‘developed to andyze the age ‘of between 1964 and 1968. The background activity is DNs, which supplies additional diagnostic informa- entirely swamped by the fission gas released from tion relating to the size and conditions of the breached elements. Figure 2 shows a typical change breach. 127-128 in cover gas activities from failure of an experimental driver. fuel element. Such failures are character- Samples of primary sodium are periodically ized by an initial increase in Xe-135m activity, taken by a sodium overflow sampler to analyze for

I I I I

--/

I I 1 I I 11 12 13 14 Date in November 1976

Figure 2. GLASS activities data failure of a metal-fueled driver element at EBR-11. n

34 the presence of 1-131, 0-137, U-235, and Pu-237. The xenon tags used at EBR-I1 consisi of about In this device, up to four IO-mL Pyrex beakers are 1 mL of a unique mixture of the stable xenon iso- filled to overflowing with sodium in an inert atmos- topes, Xe-124, Xe-126, Xe-128, and Xe-129, which phere; after five days of cooling, the beakers are is added to all elements in a given fuel assembly. removed and their contents analyzed. Throughout The tags are blended from three components: natu- the course of EBR-I1 history, plutonium has never ral xenon, natural xenon 5% enriched in Xe-124, been detected by the overflow sampler at the limit and pure Xe-128. One hundred forty tag composi- of detectability of 5 pg/g of sodium. Similarly, tions have been produced, with Xe-129/Xe-124 only twice has uranium been detected above the ratio increments of 1.25 and Xe-128/Xe-124 ratio detectability limit of 1 ng/g. This experience con- increments of 4. Figure 3 shows a typi1:al three- firms that exposure of metal or ceramic fuels to the dimensional plot of the tag nodes that correspond primary sodium has been minimal. to the experimental subassemblies in the reactor during reactor Run 11 1. With the advert of limit RBCB testing, a more sensitive device, the only distillation unit, or Whenever a failure occurred before 1977, the OLDU was installed and made operational in 1978. xenon content in about 0.5 m3 (I-20th) of the In the OLDU, a sample of about 150 g of primary argon-cover-gas volume was adsorbed on a cooled sodium is heated and volatilized. Nonvolatile charcoal trap, later released by heating, and col- impurities are left in the sample cup and may be lected for analysis on a laboratory mass splectrome- analyzed immediately, since decay of Na-24 activity ter. In this “manual mode,” sampling WEISlimited is not required. With OLDU operational, pluto- to once every 10 h and was complicated by the need nium in the primary sodium has been occasionally to prevent personal exposure. Neverthelesr,, consid- measured above its detection limit. 129 erable accuracy was obtained, and released tag volumes of 0.04 mL could be readily used for Eleven methods developed at EBR-I1 are being identification. used for the identification of fuel element failures; Table 14 lists and briefly describes them. The first five methods in the Table were used in EBR-I1 oper- In 1978, when the cover gas cleanup system ations from 1965 to 1973, when there was no (CGCS) became fully operational, a 7.5-ctn radius, accompanying release of a xenon tag. With one 60” sector, mass spectrometer was installed on line exception, these methods involved the measure- with the tag trap analysis systern.l3l Although the ment of fission gas activity in the cover gas and instrument functioned well and reliably, it had an were nonspecific. The fourth method predicted inadequate resolution for Xe-133, which is an failures on the basis of experience; it was of use for important isotope in determining subassembly fail- homogeneous groups of fuel elements such as the ures, and the instrument was also difficult to main- EBR-I1 metal driver fuel in which there was a well- tain at optimum performance. In May 19EI0, it was defined mode of failure. 30 Identification fre- replaced with a new 15-cm radium spectrometer quently involved many reactor shutdowns and that had been specially designed to have an abun- startups and 5 to 10 days of lost operating time. dant sensitivity of greater than one in 30,000 at Confirmation of an identification was always the 133 atomic mass units, to be sensitive to anlalysis of absence of short-lived fission gas activity upon very small samples, and to be more easily main- resumption of full power. tained. The performance of this new inr,trument has been extremely gratifying. It has allowed, for The remaining methods of identification in the example, the routine analysis of background sam- table were, in general, developed and used in EBR-I1 ples that contain less than 0.002-mL xencln tag in operations since 1973. The predominant method in the cover gas volume of about lo7 mL, and in the table is No. 8, the identification of a xenon tag which Xe-124 can be measured with a precision of from a failed fuel pin assembly. However, all 11 meth- 1% at a low fraction of 0.0015 of the xenon iso- ods have been used when appropriate. For example, topes. 128 This value exceeds the best sensitivity the Xe-l31/Xe-134 ratio has proven useful in distin- that could be obtained by the use of a manual guishing between failure of a uranium fissioning method by an order of magnitude, whereas the metal driver-fuel element and failure of a plutonium- 7.5-cm instrument was worse by a factor of’four or bearing experimental element. 129 five.

35 Table 14. Methods used to identify sources of fission-product release in EBR-II n

Method Purpose Advantages Disadvantages

Fission-gas Identifies suspects by Eliminates low- Usually limited volume gas release burnup suspects applicability

Ratio 134Xe/ Discriminates Identifies type and Ratio changes 133xe between metal and burnup of suspect for same element oxide affected by fuel and breach geometry

135mxe Indicates release Rapid None; occasion- of bond sodium ally overlooked

Weibull failure Ranks suspects by Predicts breach in Assumes common analysis failure probability advance; helps mode of failure; rank otherwise limited by previous equal suspects experience

Flux-tilting Narrows down sus- Easy to perform Suspect must be pects to a section adjacent to control of the core rod; only positive response is meaningful

Ratio 134Xe/ Determines burnup Eliminates sus- For small release, c 128xe level of untagged pect with too natural background element high or low contamination burnup can be significant

Ratio 131Xe/ Descriminates Uses stable high- Can be affected 134xe between metal and yield isotopes by tag in low- oxide of xenon burnup elements

Xenon tag Identifies suspects Limits choice to Exposure changes by tag one to three in tag; sometimes suspects small tag releases; contamination

Fission-gas and Discriminates sus- Ranks xenon-tag As above; also, tag volumes pects with similar suspects early tag volumes tag compositions were variable

Lift-and-hold Identifies suspects Confirms suspect Shutdown required; test by gas release subassembly in only positive fuel handling response meaningful time consuming

FUM isolation Identifies suspects Confirms suspect As above; can tolerate test by gas release subassembly at only low decay-heat operator’s level in discharged convenience; mini- subassembly mizes interference from cover-gas activity

36 90

Figure 3. Three-dimensional plot of xenon gas tag nodes for the experimental subassemblies in EBR-11 during Run 111.

The mass spectrometer is operated automatically tification and extra operating time was lost. How- by the computer of the tag trap system and is there- ever, empirical correction and, later, the result of fore required to be on-line whenever the reactor is tag exposure tests and calculations have removed at power. The instrument has operated with a mini- the uncertainty caused by composition change. mum of down time. Routine replacement of the Computer programs are now used to update com- source filament is required every five months, and positions of all xenon tags in the reactor for every the turbo molecular pumps have had slight bearing run. 131-1 33 problems. The most serious, but easily remedied Both the impurity monitoring and analysis sys- problem, has been the removal of hydrocarbons tem (IMAS) and the fuel failure monitoring system after system shutdowns. (FFMS) monitor FFTF argon cover gas. The IMAS monitors argon cover gas for process coni:rol and Although shifts in tag composition caused by in- detection of adsorber pin leakage. The FFh4S mon- core exposure were anticipated early, the magnitude itors the cover gas to detect any fuel pin, test pin, or of the shifts was not. Major changes were caused by adsorber pin leakage. Upon detection of a leaking burnout of Xe-124 by neutron capture and by pro- pin, samples of cover gas are taken to the labora- duction of Xe-128 by an (n,a)reaction of fission tory and analyzed to identify the exact fuel assem- product 1-127. During several early tag releases, bly, test assembly, or adsorber assembly that these effects caused considerable confusion in iden- contains the leaking pin.

37 The IMAS is used to perform the following Cover gas used for the reactor heat transport pri- argon monitoring and sampling functions: mary system is scanned continuously with an intrinsic germanium diode gamma detector. Fission gas iso- Continuous monitoring of concentrations topes are .monitored to detect fuel failures, and of oxygen, hydrogen, nitrogen, methane, Xe-125 is monitored to detected an absorber pin fail- and carbon monoxide in all primary argon ure when the reactor is at power. An alarm is activated cover gas systems, with on-line gas chro- in the reactor control room if an abnormal activity matographs for process control level is detected for any of these isotopes.

Monitoring concentrations of helium in The level of activity detected by the monitor var- the heat transport system (HTS) because ies from a background of 1 to 10 counts per second the presence of significant quantities of to as high as lo5 counts/s during a rapid gas helium at nominal zero reactor power indi- escape. cate an adsorber pin leak Cover gas is monitored in the cover gas monitor Providing argon to the FFMS for on-line module by passing a continuous controlled flow fuel failure monitoring through a small, liquid-nitrogen cooled, activated charcoal, sample column adjacent to the gamma Obtaining periodic bulk samples of argon sensitive germanium diode. Use of the sample cover gas from the primary and secondary column and thin detector permits accurate moni- HTS and closed loop systems (CLS) for toring under conditions of high background laboratory analysis. radiation. The technique for locating a failed fuel or The argon cover gas for the primary coolant sys- absorber assembly is based on the use of a unique tems is directly sampled from the gas equalization mixture of xenon and krypton in fuel and adsorber line connecting the reactor vessel and the sodium pins. These tag gas mixtures are loaded into the overflow vessel, and from each primary CLS surge pins at the time of fabrication. tank. A continuous closed-loop circulating flow is provided to a gas chromatograph, located in the All individual pins of a given fuel assembly con- HTS service building-south. The chromatograph tain the same unique isotopic mixture consisting of sampling head selects each gas sequentially for one standard cm3 of xenon and one standard cm3 analysis. Argon compressors circulate the sample of krypton. All individuals pins of a given absorber gas. assembly contain the same uni ue isotopic mixture consisting of two standard cm9 of xenon and two Pneumatically operated isolation valves are standard cm3 of krypton. Over 106 unique mix- automatically closed by the plant protection system tures are used for the driver, open test, the absorber (PPS) when required, to ensure containment isola- assemblies of the core. Each time a fuel or absorber tion of lines from the primary systems. pin failure is detected, cover gas samples are obtainFd with a tag gas sampling trap. Argon cover gas in the secondary system is sampled manually from the HTS secondary expansion Failure assemblies are identified by matching the tanks and the CLS secondary surge tanks, using results of a mass spectrometer analysis with previ: grab sample cylinders. ously determined analysis of all tag gas mixtures in the reactor, suitably corrected for burnup and The FFMS provides the following: background.

On-line monitoring to detect escape of fis- Unlike EBR-11, the FFTF, FFMS is not auto- sion gas from a leaking fuel assembly into mated and, it takes approximately 6 h for the segre- the cover gas, or the escape of activated tag gation of the xenon, krypton, and argon isotopes. gas from a leaking absorber assembly Few tag samples exceed 25 pCi/cm3. An ion chamber-type survey instrument (JUNO) is used to Sampling equipment and analytical proce- determine if the tag gas sample bomb exceeds the dures to locate the leaking assembly. operational safety limit of 500 mrem/h. 134 9 38 Sodium Impurity Monitoring and which decays through a beta emission with an Analysis Systems attendant release of 1.37 and 2.75 MeV gammas to u Mn-24. This activation has led to the inclusion of an intermediate coolant loop for sodium cooled Sodium Coolant. The choice of coolant for a fast systems to ensure that all radiation is confined to breeder reactor probably has a greater effect on the the primary loop. Other reasons for incluljing an overall physical plant layout and on operational intermediate loop in LMFBR systems are to protect health physics than any other design selection. the core from possible pressure surges or ,oositive Whereas the coolant strongly influences the neu- reactivity effects caused by hydrogen modleration tronic behavior of the core and provides a direct should a steam generator leak occur and result in a framework for the consideration of cladding mate- sodium water reaction, and to protect thc steam rials, the most recognizable effect of the coolant generator from radioactive corrosion and fission selection is its effect on major components such as products. pumps, steam generators, heat exchangers, shielding, and instrumentation. A final neutronics consideration concerns the reactivity effect of coolant loss during an accident. Liquid metal sodium has been selected as the The large density change associated with sodium coolant for all major fast breeder power reactor loss can cause an appreciable change in the neutron projects underway around the world. As a conse- spectrum and associated positive reactivity feed- quence, the properties of sodium and its implica- back. None of the gas cooling media exhibits such a tion for design must be given particular attention. characteristic, though a thorough safety analysis of As options having received attention to date the gas such systems does cover the very remote possibility coolants, helium and steam, have been given lim- of water entering the core as a result of a major ited attention, as have other metal coolants. break or depressurization. Under such conditions, similar variations in reactivity could occur. The very high power density in an LMFBR core of %400 wattdliter compared to %lo0 wattdliter In order for a reactor coolant to perform its heat for an LWR places very stringent requirements on removal function, a fraction of the total power output the heat transfer properties of a coolant candidate. must be used to pump the coolant through the core. A large heat transfer coefficient is desirable for Hence, it is of interest to compare pumping require- optimum heat removal. Sodium is superior from ments for the various coolant candidates. For a typi- the standpoint of heat transfer characteristics, cal size plant, sodium coolant requires the least, and though helium and steam can be driven hard helium requires the most, pumping power. Other liq- enough, given sufficient driving pressure, to uid metals are written off, largely because of their accomplish the required tasks. much larger pumping requirements. 135 Sodium is quite compatible with the preferred Whereas any coolant has some moderating effect cladding candidates. However, because sodium has on the neutron spectrum, the degree of moderation been chosen as the coolant, much research has been taking place is proportional to the atomic mass and directed to every facet of its in-core environment, in the density. Sodium, with a mass of 23, is clearly order to determine any long-term deleterious heavier than either helium or steam, but when the effects. Summarized below are the main categories density effect is taken into account, the helium- where effort has been focused to isolate pxsible cooled reactor yields the hardest spectrum; steam problem areas. yields the softest spectrum because of its hydrogen content; and sodium provides an intermediate spec- The principal metallic elements in the clad- trum. Because of the spectrum differences, helium ding, such as iron, chromium and nickel, has a slight intrinsic advantage in terms of achieva- are very slowly dissolved in sodium from ble breeding ratios. Another neutronic consider- the hot core region and deposited in the ation is the effect of coolant activation. Helium, cooler areas. This long-term uniform having a neutron capture cross section of essen- attack is called cladding thinning or wast- tially zero, is not activated. Steam becomes some- age. Although this amounts to only tens of what radioactive by the neutron activation of micrometers/y at 700°C, it must bl: con- oxygen, but the impact on design is minor. Sodium sidered in the overall load-bearing capacity activation, however, attains a high degree of short of the cladding, which is stressed by inter- half-life activation. Na-23 is activated to Na-24, nal fission gas pressure.

39 Each of the principal metallic elements port media for the general corrosion of mentioned above has a different dissolu- cladding, the selective leaching of metallic tion rate in sodium. Hence, selective leach- elements, the transport and deposition of Q ing occurs, and the composition of the dissolved metals and radioactive species, steel at the outer surface is eventually and the transport of carbon, samples of altered. The effect, though small, is nor- sodium are usually extracted regularly and mally deleterious, since chromium and analyzed during normal operations at nickel are removed faster than iron. Some LMFBRs. On-line instruments have been of the more significant problems arising developed to monitor sodium purity for from the dissolution of minor constituents both the primary and secondary systems. or activation products from the steel in the In addition to determining oxygen and car- core are discussed below. bon content, it is useful to monitor hydrogen levels, since this provides an indication The dissolved metal deposited in the colder of the tritium level in the cover gas of the regions of the intermediate heat exchanger primary system or steam generator leaks in can lead to high heat transfer resistance or the secondary system. reduced heat exchanger efficiency. The A principal advantage of sodium, beyond its deposits can become thick enough that excellent heat transfer properties, is that it does not coolant driving pressure must be increased require pressurization to prevent coolant boiling. in order to maintain desired flow. Some This feature contrasts sharply with the thick-walled long-term silicon de osits have been high-pressure systems required for helium or observed at EBR-11. 139The silicon, origi- steam-cooled systems. Table 15 shows the pros and nating as a minor constituent in the steel, cons of sodium coolant and other items that bear forms a compound with sodium which, if on coolant selection. precipitated in core assembly coolant channels, can influence pumping power or , Sodium Cold Trap Technology. In the operation of local flow conditions. LMFBRs, it is necessary to maintain control of the amount of impurities in the liquid metal coolant sys- Though small in quantity, sodium trans- tem. The health physics department usually has the port and deposition of manganese and responsibility for sodium quality. Control of the level cobalt in the intermediate heat exchanger of such impurities as hydrogen and oxygen is required and other primary system components, because of in-core metalic elements. Oxygen is can cause high enough radiation fields to sodium enhances the corrosion rate of austenitic affect routine maintenance. stainless steel and the rate of embrittlement of refractory metals. Hydrogen may cause embrittlement of refractory metals. Both of these impurities form com- In LMFBR secondary sodium systems that pounds with sodium whose solubility in sodium incorporate dissimilar materials in differ- decreases with decreasing temperature, leading to pre- ent regions of the circuit, nonmetallic ele- cipitation and deposition of the compounds. All of ments such as carbon and nitrogen are these effects can increase maintenance requirements known to migrate as a result of chemical at the plant and thereby increase the radiation expo- activity differences. Ferritic steel is suscep- sure to plant personnel. Many methods of treatment tible to decarbonization when exposed to of the sodium to remove the impurities has been pro- high-temperature sodium because of its posed and investigated. Among them are filtration, inherently high carbon activity and large settling, centrifugation, distillation, chemical reac- carbon defusion coefficient. The carbon tion, soluble gettering, and cold trapping. 137 loss from this material in a sodium environment leads to a significant reduction in The most widely used method for control of the tensile strength and stress rupture life. Fur- hydrogen and oxygen impurities in operating thermore, the carbon transferred to the LMFBRs has been the use of cold traps because of sodium can carbonize the austenitic stain- their simplicity, economy of operation, and relative less steel used in the intermediate heat effectiveness. Optimization of cold trap design and exchangers and piping in the same circuit, operation can increase capacity and lifetime, with a resultant loss in ductility of the thereby improving the overall economics of the stainless steel. Because sodium is the trans- system.

40 Table 15. Intercomparison of FBR coolants

Coolant Advantages Disadvantages

Sodium -Excellent heat transport -Radioactivity (intermediate, properties sodium loop employed) -Low pressure system -Unfavorable coolant reactivity void coefficient -Low pumping power requirements -Chemical reactions with ail, and water -Lowest fuel cladding -Nonvisible refueling procedure temperature -Potentially high breeding -Solid at room temperature ratio -Inherent emergency cooling -Maintenance on primary sq'stem of fuel impeded by radioactivity

Helium -No intermediate loop -High pressure system -Coolant not activated -High pumping power require- -Potentially high breeding ments ratio -Cladding roughening required -Visible refueling/maintenance -Emergency cooling provisicins -Minimal void coefficient not established -Most compatible with materials -Unproven high power density -Utilization of thermal GCR capability technology -Lack of FBR technology -Potential for vented fuel -Gas leakage difficult to -Potential for direct cycle control -Flooding with H20 tolerable -High performance demand:; on pumps and valves

Steam -Direct cycle -High pressure system -Visible refueling/maintenance -High pumping power -Industrial capability -Cladding corrosion available for components -Lack of FBR technology -Minimum chemical reactions -Emergency cooling provisicn -Fluid at room temperature not established -Low breeding ratio -Fission product carryover to turbine. -Unfavorable coolant reactility coefficient.

Sodium in the LMFBR intermediate heat trans- Measurements of this corrosion rate on full-scale port system is subject to significant contamination reactor systems have shown that the hydrogen by hydrogen derived from normal corrosion of the source is very large (1.8 x g H/~m~-s).l~~ steam generator tubes. The Fe-2- 1/4, Cr- 1 Mo steel The quantity of hydrogen thus entering the sodium used in the steam generators oxidizes slowly in the in a typical 500-MWe LMFBR calculates to be steam environment, releasing nascent hydrogen, about 25 kg/y. It is important that steady state which defuses through the steel into the hydrogen levels be maintained below 200 ppb in the sodium. 138 Most of this hydrogen defuses through intermediate heat transport system sodium to per- the tubes into the sodium coolant and must be con- mit early detection of possibly small steam genera- tinuously removed. tor leaks that would introduce water into the

41 sodium. The capability of a cold trap may be such as oxides or hydrides remain soluble in the n exceeded in a short time because of the increase in sodium. Under proper conditions, these impurities the pressure drop across the cold trap. Estimates precipitate on the surfaces of the mesh packing for indicate that a normal cold trap, when used to walls of the cold trap. remove hydrogen from the sodium, has an expected lifetime of 12 to 18 months at full power opera- Early designs of cold traps have featured a nar- tion. Replacement of intermediate heat trans- row, unpacked or mesh-pack annulus with a large port system cold traps on an annual to 18-month mesh pack central region. Examples are the Mine schedule would be prohibitively expensive and Safety Applications (MSA), Fermi, and EBR-I1 would result in large quantities of tritium-bearing cold traps. 141-143 Analysis of two of these traps waste sodium in the spent cold trap. indicates some deposition in the annulus but high utilization at the inlet to the center region. 144 The Figure 4 schematically depicts a typical forced- reason for this behavior is thought to be that as the circulation flow-through type cold trap. The sodium passing through the unpacked annulus sodium flow enters at the top, passes down through becomes cool to below the saturation limit for the an annual region, laterally through an unpacked impurity level present, the impurities precipitate bottom region, up through the center region, and out on the mesh. This is because the energy out of the cold trap to a heat exchanger, and then required to nucleate on a surface is less than the returns to the reactor system. In the coolant chan- energy required to form a particle in the solution. nel, NaK flows counter-current to the sodium flow There is also the possibility of micron size particles in the annulus. Cold-trap operation is based on the of NaH being in the sodium, but the coarseness of dependence of impurities solubility with tempera- the mesh does not allow it to serve as an effective ture; the lower the temperature, the less impurities filter to trap any particles that are in the sodium.

Sodium Sodium out in .

II IbI

I Packing (mesh)

Coolant channel

Figure 4. Schematic of a sodium cold trap for MASCOT calculations. n

42 -......

The achieved capacity of cold traps has histori- required by operational testing prior to installation cally been low. Typically, only 10% of the total in a system. Because the performance of the cold capacity of the cold trap is used before the pressure trap is closely tied to inlet sodium temperature and drop across the trap is too great, necessitating impurity concentrations, the manner of control of replacement of the trap. The increase in pressure this temperature and measurement of the impurity drop is caused by local depositions of impurities, concentration must be addressed in the design of which occupy up to 90% of the available flow vol- the cold trap system. ume. A goal of 40 to 50% average use of cold packing capacity is theoretically attainable. 145 Though the principal role of the cold lrap in the primary circuit is to control the concentration of Parameters judged to be important factors in dissolved oxygen impurity and, in the intermediate influencing the lifetime of a sodium cold trap are as heat transport system, to control the concentration follows: of hydrogen impurities, a number of radioactive impurities have been observed to be cold trapped Sodium inlet flow rate also. Cold traps are not designed with this inten- tion, and in the case of many such species it is not Sodium inlet temperature clear that the mechanism of cold trapping is precipitation from solution. The parameters that affect Minimum sodium temperature in the cold the phenomena of crystal nucleation, growth, and trap mass transfer, which are fundamental to cold trap performance, have been defined. 146 The perform- Impurity present in the inlet sodium and ance of a cold trap in relationship to a given radio- its concentration level active species cannot be described simply, in terms of precipitation, and will be markedly sensitive to Mesh packing density design and operating parameters. In addition to providing a site for crystal nucleation and growth, a . Mesh packing wire diameter cold trap is characterized by high residence time of the sodium, heat exchange, and flow conditions, Location of the mesh packing which may be conducive to paiticulate deposition-depending on the design, on the high Length-to-diameter ratio of mesh packing surface areas of steel in the trap internals and diso- section dium monoxide at low temperatures compared to the remainder of the primary circuit on which dep- Frontal flow area ratio (annulus/center osition may occur preferentially, and on the struc- section). tures conducive to the filtration of particulate material. The last aspect, in particular, may be sen- The length-to-diameter ratio is the ratio of the sitive to operating mode; it has been suggested that length of the annulus region to the outer diameter the PFR cold trap will only become an zffective of the annulus region. The flow area ratio is the filtration unit for particles below 100 microns ratio of the total frontal area of the annulus region when the trap has a considerable inventory of diso- to the total frontal area of the center region. dium monoxide-type impurity through the mesh packing. 147 There are many possible approaches to the choice of a system design for large-scale LMFBRs. The The potential role of cold trapping as a means of number and size of secondary systems on the reactor removing radioactive products from the coldant of will guide the selection of the number and, to a LMFBRs has not been explored fully. Thcre have degree, the size of the intermediate heat transport sys- been few systematic studies of cold trapping per- tem cold traps. General guidelines given in Table 16 formance in operating reactors, possibly as a result will aid the designer in the area of intermediate heat of the difficulty of carrying out such analysis; the transport system cold trap design. These guidelines information available is fragmentary and does not apply to forced circulation flow through type IHTS lead itself to generalization. Furthermore, ihe per- cold traps, calculated using the MASCm code.145 formance of the cold trap in removing radiioactive products will be intrinsically related to the deposi- It should be emphasized that verification of the tion behavior of these species in the remainder of performance of a proposed cold trap design will be the primary circuit, and this deposition bzhavior

43 Table 16. General guidelines for IHTS cold trap design n

Item Guideline

General

(H) Impurity source rate Calculate based on heat-exchanger size, design, and materials

Utilization, vol% > 38

Efficiency, Vo > 75

Lifetime, mo > 40

Design

L/D ratio 1 to 1.5

FAR 4 to 8

Mesh

Wire diameter, cm ~0.028

Density, g/cm3 0.32-0.4

Location Annulus only

Size Range, L 3000-7000

Recommended pressure drop, Pa -4.8 x 104

Operating

Maximum (H) impurity level, ppm 0.2

Temperature, "C

Inlet sodium At or slightly (1-2°C) above the saturation temperature for the impurity level in the inlet sodium

Minimum sodium 114-125°C

Inlet NaK coolant (range) 6545°C

Flow rate, LIS

Inlet sodium To yield a residence time of 10-15 min

Inlet NaK coolanta Equal to the sodium coolant rate established above a. If gas coolant is used, then the flow rate is that required to remove the heat load and obtain the minimum trap temperature. n ~~ ...... --

will vary from system to system, as described previ- The efficiency, P, was found to be unity, or 100%. ously. Table 17 lists radionuclides, excluding trit- Hence, all iodine released to the circuit was cold ium and those of sodium, detected in LMFBR cold trapped after the required number of passes. In traps. contrast, the maximum proportion of the (3-137 inventory of the primary circuit observed to be cold At Phenix, the only release of fission products to trapped was 50%.150-151 date is thought to have occurred as a consequence of a gas leak, rather than a fuel fai1~re.l~~The More than 90% of the inventory of both iodine cesium levels observed were close to the minimum and cesium isotopes in the BR-5 primary circuit was detection limits and not viewed as significant. Data cold trapped.152 In terms of the proportions cold indicate that cold trapping of radionuclides is trapped, the cold trap was found to be much less inadequate for meaningful analysis at present. 149 effective for Zr/Nb-95 and Ba/La-140. 152

At the RAPSODIE LMFBR, the efficiency of The BOR-60 cold trap system was specifically cold trapping of 1-13 1 has been investigated using aimed at assessing the efficiency of removing radio- data on the kinetics of 1-13 1 release from failed fuel active impurities as a function of design anld oper- and concentration changes in the primary sodium, ating parameters. Following decay of short-lived defined as 1-131 and Te-129m isotopes, ~150Ci of long-lived isotopes had accumulated in the cold trap by p=-‘e - ‘s June 1974. The principal contributors to this activity were Cs-137 (130 Ci), Cs-134 (9 Ci), and Co-60 Ce (2.5 Ci). where Purification of the sodium coolant achieved by P = the efficiency of cold trapping cold trapping was quantified in two ways, as presented in Table 18. A purification coefficient was Ce = the concentration of 1-131 at the evaluated, where entrance to cold trap K=A (A +A) Cs = the concentration at the exit. ct c ct

Table 17. Radionuclides detected in primary circuit cold traps of FMFBRs

Reactor Reference Radionuclide

Phenix 42 Mn-54, Co-58, Co-60, Zn-65, Sb-124, Cs-134, CS-136, CS-137

Rapsodie 150 1-131, CS-137

BR-5 152 Zr-Nb-95, 1-131, 1-133, 1-135, (3-136, Cs-137, Ba-La- 140

BOR-60 161 Mn-54, Co-60, Zn-65, Zr-Nb-95, Ag-1 lorn, Sb-124, Sb-125,1-131, CS-134, CS-137

EBR-I1 154, 155 Mn-54, CO-60, Zn-65, Sb-124, Sb-125, 1-1 31, (3-134, (3-137

DFR (Na/K coolant) 40 Cr-51, Mn-54, Zn-65, Zr-Nb-95, Sb-124, Sb-125, Te-l29m, 1-131, Cs-134, (3-137, Ba-La-140, Cs-141, Ce-144

45 Table 18. Purification coefficients for the BOR-60 cold-trapa n

Purification Coefficient K K, Kb

Na-22 0.054 1 1 CO-60 0.21 0 88 Zn-65 0.86 6.6 19 Ag-l10m 0.043 0.83 0.75

Sb-124 0.22 4.4 6.2 1-131 0.99 1.5 104 2.1 104 (3-134 0.32 4.0 15 CS-136 0.30 5.2 15 CS-137 0.38 4.7 26

a. Taken from Reference 1-153. where virtually no further cold trapping of cesium was observed after 1973. The proportion of the total Act = total activity of an isotope in the cold cesium inventory in the cold trap fell correspond- trap ingly from the values shown in Table 18 to ~0.2. . A, = total activity of an isotope in the The distribution of isotopes in the cold trap was whole of the remainder of the circuit. also investigated and found to be far from uniform. The concentrations of all isotopes investigated This coefficient, K, is a function of plant design came to a maximum at the entrance to the first fil- parameters. For example, the value of K for Na-22 tration zone. is equal to the ratio of the volumes of the cold trap and the primary circuit; for BOR-60, this ratio is Mn-54 was not cold trapped as readily as Co-60; 0.054. A specific purification coefficient was also Zr/Nb-95 was not cold trapped to any significant evaluated, which is independent of parameters such extent; and Ba/La-140 did not appear to be cold as cold trap volume (defined as the ratio of the spe- trapped. cific activity in the settling zone), or the zone of filtration in the cold trap to the average specific Cold trapping at EBR-I1 is considered to occur activity of the same isotope taken over the circuit either by a precipitation mechanism or by a (K,and Kb, respectively). These coefficients were segregation-adsorption mechanism, the latter found to be fairly constant for measurements over mechanism applying to all fission and activation the period 1972-1973 for BOR-60 and were found products. The segregation adsorption mechanism to be independent of parameters such as cold trap appears to be less effective than the precipitation volume. mechanism for the removal of hydrogen and oxygen in the EBR-I1 cold traps.154-156 Nearly all the iodine in the primary circuit is retained in the cold trap. The kinetics of cold trap- Studies conducted at EBR-I1 have demonstrated ping of 1-131 were also investigated. An extraction the following: constant of 0.12 h-l was evaluated, corresponding to almost complete removal of I- 131 from the main 1-131 was not removed effectively after a circuit after 24 h of operation of the cold trap.153 clad failure in July 1975; radioactive decay of the 1-131 inventory was the major The performance of the cold trap with respect to removal mechanism. This observation cesium showed a marked variation with time, and contrasts with other experiments at EBR-I1

46 where lowering of the sodium temperature cies remain largely unresolved. Before assessing the produces a quantitative segregation of contribution that might be made by cold trapping to 1-131 to metal surfaces. This segregation the removal of radioactive products from LMFBR process provides the basis for an analytical primary circuits, it is desirable to consider the mecha- method for 1-131 used at EBR-11. nism of cold trapping of specific radionuclides.

During the first year of operation of the cold Appreciable segregation of Cs-137 in cold traps trap (1968-1969), 30% of the total Cs-137 has been reported under certain condition:. How- inventory was located in the cold trap. Sub- ever, the conditions that will optimize cold trapping sequently, the absolute quantity of Cs-137 in of Cs-137 have not been identified. Varialions in the cold trap has remained effectively con- the nature of steel surfaces, flow conditions, and stant, and further cold trapping of the iso- the coolant chemistry within the cold trap may all tope does not appear to have occurred. account for the observable differences in cold trap efficiency at LMFBRs. A number of workers have The cold trap is extremely effective for conducted experiments in order to improve the Mn-54 removal, removing ~60%per pass. understanding of cesium deposition behavior in However, the benefits derived from effi- sodium/stainless steel systems. Classically, ii distri- ciency in this respect are dubious for con- bution coefficient, K, is obtained to describe the tinuously produced corrosion products distribution of cesium between its deposited form such as Mn-54. This removal in the cold on a steel surface and its dissolved stage in siDdium, trap will not influence deposition on com- where ponents upstream; most of the Mn-54 will deposit in the primary circuit and not enter atoms Cs/cm -L steel surface K= the cold trap circuit. -3 (3) atoms Cs/cm Na Generally, radioactivity was not uniformly dis- It has been demonstrated for a variety of steel tributed in the cold trap. The level of activity was at surfaces, cesium concentrations, and oxygen levels a maximum at a point close to the inlet for the in sodium, that the log of K is inversely propor- external cooling (employing NaK alloy) where most tional to the absolute temperature. 157-165 ]Hence, precipitation is thought to occur. cesium will deposit preferentially on the cold steel surfaces in the cold trap. However, even taking the The cold trap at DFR was found to be highly strongest temperature dependence reported does effective for the removal of 1-131 and Tc-l29m; not offer a practical route for significant (cesium Ba/La-140 also segregated markedly to the cold removal from the coolant of an operating LR4FBR, trap. However, the efficiency of the cold trap in since the much greater surface area of the reactor removing Cs- 137 was deemed to be disappointingly will result in deposition of too high a proportion of low. The behavior of antimony was variable, and the cesium outside the cold trap.160 This conclu- association with fuel particles was postulated. sion is similar to that reached in the context of the BOR-60 cold trap.161 The ease with which 1-131 in the cold trap was removed by steam cleaning suggested it to be The distribution coefficient, K, has been found present as precipitated sodium iodine, whereas the to strongly depend on the nature of the steel acid cleaning required to remove Te-129m was surface. 57-160 Pretreatments such as polishing or taken to indicate some weak chemical binding to preexposure to sodium at high temperature l(about - -- . the surface of the steel mesh cold trap basket. The 600°C) resulted in smaller values for K. This is 4” - decontamination behavior of Ba/La-140 (and the ascribed to the removal of oxide films, at the steel species Zr/Nb-95, Ce-141, and Ce-144) appeared surface, on which cesium deposition occur:, more indicative of fine surface oxide deposition. How- favorably than on ~tee1.I~~Nickel and alumina ever, cesium appeared strongly bound to the steel of have been shown to be more efficient cold trap

the cold trap, and penetration of the steel was pos- packing materials for cesium removal. :i79 159 tulated to explain the resistance of the cesium to Investigation has shown that disodium monoxide decontamination procedures. adsorbed cesium more effectively than steel. Hence, enhanced cesium removal could be Observations on the cold trapping behavior of obtained in a primary circuit cold trap in which radionuclides are often in conflict, and the discrepan- oxygen is cold trapped continually. Such a sy:,tem is

47 penalized by the intrinsic requirement of periodic adsorption compound that subsequently yielded a regeneration or replacement of the cold trap and by more stable intercalation compound. The effi- interactions with the aspects of coolant chemistry. ciency of trapping was further proved by using a Investigations of this effect have proven inconclu- high surface area charcoal (gas adsorption char- sive.161 coal) that was thought to enhance the formation of the intercalation compound. Short-term experi- The dissolved oxygen in the sodium influences ments showed no adverse effects of preexposure to cesium deposition on steel by the interaction of a sodium environment on trapping performance. oxygen with the steel 166 Cesium may The use of carbonaceous material for trapping be incorporated in corrosion products at steel sur- cesium in the sodium coolant of an LMFBR pri- face~.~~~Some evidence suggests that there is little mary circuit might give rise to unacceptable levels correlation between the enhancement of cesium of carbon on steel components, as a consequence cold trap ing and the oxygen level of the of carbon transfer. However, many forms of pure sodium. 16f-164 carbon are inefficient carbon sources for steel carburization in sodium, and the performance of high Cesium adsorption at low temperatures may be surface area graphite for charcoal in a sodium/ controlled by interactions with nonmetallic impuri- stainless steel system appears worthy of further ties, and the strongest interaction occurs with investigation. 185 Carbon trapped materials have hydrogen. 167 A phase distribution coefficient, D, functioned efficiently in the cover gas space above has been evaluated, where the sodium, though in this mode the occurrence of steel carburization may not be precluded. atoms Cs/g hydrogen in deposit D= atoms Cs/g sodium In the EBR-I1 primary circuit sodium, the performance of reticulated vitreous carbon (RVC), a At 120"C, D has a value of 8.5 x lo-? 'This highly porous foam-like glassy carbon manufac- seems to indicate that hydrogen additions to the tured by Chemotronics, International, has been primary circuit, with subsequent cold trapping, will investigated for use in removing cesium from enhance the cesium removal. sodium. 185 Favorable results have been obtained in the laboratory, and reactor tests conducted on Although distribution coefficient measurements graphite indicate that RVC is preferred for use in for a range of concentrations of Cs-137 in sodium the reactor in view of its greater surface-to-volume have been made,159 there remains a lack of long-term ratio. RVC specimens were exposed at various tem- data obtained under controlled, representative condi- peratures to flowing sodium taken from the pri- tions on the cold trapping behavior of cesium. The mary tank, which was maintained at 371°C and effects of indigenous levels of nonactive cesium impu- contained 0.3 pCi Cs-l37/g sodium throughout. rity in the original sodium charge have been largely Over the temperature range of 177-371°C, it was ignored. The nature of active cesium trapping sites found that RVC traps Cs-137 according to a rela- remains obscure. Experience from BOR-60 and EBR- tionship given as I1 suggests that such sites might have to be produced continually for regeneration of the cold traps. Log C = 3,00O/T - 9.889, Evidence indicates that cesium adsorption on surfaces is reversible and highly sensitive to a num- where ber of variables that may not be controlled suitably during LMFBR operation. These findings suggest C = C-137 saturation (Ci/cm2) that a trap for cesium should bind the cesium more strongly than the forces involved in adsorption, and T = temperature (K). that the nature of the bond should be insensitive to other concomitant phenomena. The trapping mechanism remains obscure; if the active surface involved is estimated correctly, then The stabilities of cesium-graphite intercalation 75 molecular layers of (3-137 will be present at the compounds and the sorption of cesium by graphite RVC surface at 177°C. Furthermore, the material is in liquid sodium have been investigated, 83-184 said to be effective for Cs-134 and presumably for and it is concluded from them that cesium was nonactive cesium. Therefore, it seems likely that a trapped in two stages, the first involving a surface mechanism other than adsorption applies.

48 The RVC for EBR-I1 that will handle the present behavior. Assuming small particle sizes of 1t:ss than Cs-137 inventory and the further release caused by 1 micron, the deposition behavior of Zr/Nb-95 Grs the run-beyond-cladding-breach program will and Ba/La-140 from sodium on steel surfaces operate at 193°C and contain m3 RVC, giving might be described adequately in terms of liquid- an effective surface area of 320 m2. 146 The cesium phase mass transfer control. However, the strong trapped is placed within the cold trap and consists temperature dependence of Ba/La- 140 deposition of an economizer, RVC, and crystallizer, and is sur- in BOR-160, implies the use of various liquid-phase rounded by a 4-in. lead shield. The normal sodium mass transfer coefficients at different tempera- flow through this cesium trap is 12-15 gal/min, tures. 170 These observations suggest that adsorp- which at this flow takes approximately four to five tion of Ba/La-140 on steel surfaces is more days to clean the entire sodium system at EBR-11. favorable at low temperatures; a similar observa- When the cesium trap was first in place, the total tion has been made for strontium.171 In a practical cesium inventory in the sodium was in the range of system, these higher adsorptions at low tempera- 120-150 Ci. Presently, the maximum cesium inven- tures will allow only a small proportion of the Ba/ tory in the sodium is 10 Ci and is normally in the La-140 inventory to be cold trapped. This is in good range of 2-3 Ci.136 agreement with observations on BR-5 and BOR-60, and with experimental data.152,171,174 Halogens released into liquid sodium are converted to sodium halides, which have a well-defined Analysis of sodium/stainless steel system:. shows solubility in sodium, and have been measured that Co-60 deposition experimentally. 68 The solubility of sodium iodide shows a very strong temperature dependence, but Is approximately proportional to sodium precipitation from solution of sodium iodide at low velocity at all temperatures temperature in the cold trap of a primary circuit cannot be proposed as a method for 1-13 1 removal, Decreases with downstream distant': since iodine levels above the solubility limit, even at cold trap temperatures, are unlikely to be attained. Shows little variation with temperature above that from downstream behav- Generally, iodine appears to be amenable to cold ior.172,175 trapping, but by adsorption rather than precipitation. The temperature dependence of a distribution In contrast, analysis of sodium/stainless steel coefficient, K, defined for the case of cesium, has systems has shown that Mn-54 deposition been evaluated for iodine by several investiga-

tors. 162-164, 168y 169 There is some evidence that K Shows no velocity dependence at high tem- is concentration-dependent. 168-169 The increase peratures in values of K for decreasing temperature is sufficiently great that adsorption on steel surfaces in the Increases with reducing temperature, espe- cold trap will be significant in removing 1-131 from cially approaching the minimum ternpera-

the primary circuit of a reactor. 1679 169 The cold ture in the system trapping efficiency will be further increased when disodium monoxide or sodium hydride are being Shows a velocity dependence at low precipitated in the cold trap because of a strong temperatures. 1729 174-176 interaction between iodine and the hydride deposits. 169 The observation from EBR-I1 con- Co-60 deposition behavior is found to be contrasts with this conclusion, and it must be postu- trolled by laminar boundary layer, and Mn-5 4 dep- lated that the surfaces in the EBR-I1 cold trap were osition behavior is controlled by solid state deactivated, with respect to iodine adsorption, by diffusion at high temperatures and by bulk deposi- the concomitant processes. 54-155 tion at low temperatures. These findings indicate that cold trapping will be ineffective for cobalt iso- A number of important fission products will be topes, but that Mn-54 might be cold trapped quite released into the sodium in the form of stable effectively on cold steel surfaces. Such concliisions oxides. These include Zr/Nb-95 and Ba/La-40. are borne out by observations on RAPSODIE, Examination of the cold trapping behavior of these EBR-11, and the KNK reactor, but contradicied by isotopes is effectively a treatment of small particle data from ~0~-60.154,1659171,187-188

49 A further phenomenon, which might enhance after 1 year of operation.181 Still wider spacings cold trapping of these transition metal species, is will have to be incorporated into this type of trap formation of complex oxides as a result of interac- before the pressure drop would be acceptable in a Q tion between the metal in solution and the oxide at commercial LMFBR. saturation in the cold trap sodium. The nickel used in fabrication of the traps has a Analysis of the controlling atomic diffusion very low cobalt impurity level (<5 ppm) to avoid processes confirms that enhanced diffusion of adding significantly to the CO-60 inventory in the some elements into nickel should occur. 179Hence, reactor. Moreover, the corrosion rate of the nickel use of a trap fabricated in nickel mesh, rather than in the traps was not found excessive. The finding is stainless steel, and placed at the core outlet, should based on weight change data, however, and estima- prove effective for Mn-54. Possible difficulties are tion of corrosion losses by this method for a mate- the higher rate of bulk corrosion of nickel, com- rial that shows a strong tendency to pick up other pared with stainless steel, and the modification of materials (particularly iron) may be misleading. 82 the nickel surface by concomitant mass transfer of Release of nickel in subsequent deposition, and other elements. The use of a nickel substrate does cooler parts of the reactor circuit, may lead to sec- not greatly influence cobalt deposition. 1433 177 ondary trapping of Mn-54, in the long term, at the surface of the components that the traps were In screening tests on a wide variety of materials, intended to protect. mostly pure metals and alloys, it has been shown that pure nickel is the most effective getter for Trai ni ng Mn-54 of the materials tested.180 A number of The development of LMFBRs has required a vast traps using nickel foil have been fabricated and design and experimental program in the field of tested. These are made by rolling the foil onto a sodium technology. Training and qualification pro- cylindrical form, with the spacing between succes- grams in sodium technology, sodium systems, and sive layers of the foil being determined by a design components particular to LMFBRs are needed. embossed in the foil. The programs should address the selection, training, and retention of all individuals involved in In application of this type of trap to a reactor, operational health physics. In Europe, LMFBR traps are situated at the end of subassemblies, operating experience demonstrated the specific downstream of corroding core components. Tests needs for sodium and LMFBR simulator-based on prototypical designs have been carried out training courses. This section is essentially based accordingly. on information gathered from the training programs set up in France by the Commissariat a Initially, traps with very narrow interlaying spac- I’Energie Atomique for the European LMFBR ing IO maximize surface area were tested and found community and the sodium technology course of 82-92% efficient for Mn-54 removal in terms of Atomics International. activity retained in the trap, compared with that released to the remainder of the downstream test Personnel such as reactor operators, radioactive circuit. In order to reduce a very high pressure drop waste operators, supervisors, radiological engi- associated with this design, traps with wider inter- neers, ALARA engineers, health physics techni- layer spacings in different embossed patterns have cians and supervisors, radiation contamination been tested. Traps with twice the initial spacing control engineers, chemists and chemical techni- have been shown to have equally high efficiency for cians, and other technical support should have Mn-54 removal, suggesting that trap surface area is extensive sodium training courses. These courses not of overriding importance, and that lower sur- should cover the following: face area may be compensated by the increased sodium half length and turbulence. The efficiency The chemical, physical, and nuclear prop- of Co-60 removal is reduced on going to lower sur- erties of sodium. face area, as would be expected. The solubilities of metals, gases, and non- The pressure drop associated with this design is metals in liquid sodium, and factors that acceptable in an EBR-I1 subassembly channel, and affect these solubilities, and the reactions subsequent to satisfactory performance in hydrau- of sodium with inorganic, ceramic, and lic tests. An EBR-I1 test was successfully finished organic materials. n

50 ...... -. .. . - ...... -. .-

The mechanisms of corrosion and mass - The critical temperature, pressure, transfer phenomena in sodium, their limit- and volume of sodium. ing factors, their measurement techniques, the corrosion behavior of metals in - The electrical resistivity of sodlum as a sodium, carbon, nitrogen, hydrogen, and function of temperature. oxygen transport in sodium, corrosion product deposition, and fission product - The density of liquid and solid sodium transport. as a function of temperature. The purification, sampling, analysis, and the analytical instrumentation used on a function of temperature and pressure. sodium systems.

A study of the various process systems that - The isothermal compressibility of liquid make up an operating LMFBR, such as the sodium as a function of pressurc:. sodium pumps, intermediate heat exchangers, valves, steam generators, air- - The thermal conductivity and diffu- cooled heat dump exchangers, cold traps, sivity of sodium as a function of vapor traps, and plugging meters. temperature.

The mechanical properties of sodium sys- - The kinematic and dynamic viscosity, tems, including their high-temperature surface tension, and Prandtl nuinber of properties, how these properties are deter- sodium as a function of temperature. mined, how they are interpreted, and how they are used as a basis for design, and the The most important nuclear properties of radiation effects on materials, and the sodium are the neutron capture cross section of metallurgical phenomena such as stress sodium reactions as a function of neutron energy, corrosion cracking. the associated neutron reaction, the products formed, their half-lives, and emissions. The chemi- Sodium instrumentation and electrical sys- cal interactions that should be studied include the tems such as flowmeters, level instru- solubility of metals, gases, and nonmetals in liquid ments, pressure instruments, temperature sodium, and the factors affecting their soluliilities. instruments, and leak detectors.

The operational considerations in the use The reactive nature of sodium and the r a d’ioac- of sodium systems such as maintenance, tive content of the coolant demand special si.udy of storage and handling, the potential hazards involved in operating sodium systems. The subjects of particular concern include Fire safety, and sodium-water reactions. the following: The significant chemical and physical properties of sodium that should be stud- The characteristics of sodium po801and ied due to their importance in the safe spray fires, fire extinguishing methods, design and operation of LMFBR systems and detector types for both sodium leaks includes: and fires

- Heat capacity of liquid and solid Sodium water reactions, their detlxtion, sodium as a function of temperature. their effects, and results of studies and tests carried out in this field. - The vapor pressure of sodium and its variation with temperature. The effective, safe, and economical use of sodium systems requires special knowledge for - The heat of vaporization, the latent operation and maintenance. Repair and modifica- heat of vaporization, the vapor heat tion of a sodium system require special consider- capacity, and the relative contribu- ations, such as checking whether the system is tions of sodium monomers and dimers completely drained, as well as checking the temper- as a function of temperature. ature distribution, the amount of oxide or other

51 impurities present, and the need for an inert gas Tripping of one or more pumps with or cover. Major work on a sodium system involving without inertia acceleration or decelera- the cutting and replacing of pipe will require train- tion, in both the primary and secondary ing distinctive from that for LWRs. circuits Tripping of the valves actuating heat LMFBR simulators need to be designed for exchangers in the secondary circuit studies of operating rules, checking of operating procedures, and the training of reactor operators. Cold startup and the use of the procedures The LMFBR simulators are able to model the fol- allowing the passage from one power set lowing components: up to another Normal and emergency procedures for e The core, including all types of subassem- reactor shutdown. blies, and a neutron kinetics systems, which takes into account the majority of The sodium technology and LMFBR thermal feedbacks simulator-based courses described above are of the type taught at Cadarache, France, by the Commis- sariat a ['EnergieAtomique for training personnel e The primary sodium circuit, including all from Super Phenix. The Nersa Company, incorpo- types of pumps rating utilities from France, the Federal Republic of Germany, and Italy, also has its personnel trained e Intermediate heat exchangers at Cadarache. The training needs were originally defined for model testing and setting up of reactor 0 Secondary sodium circuits component prototypes, operation of test facilities, and the commissioning and operation of the Rap- 0 Steam generators sodie and Phenix Reactors. Since 1975, the school has taught designers, manufacturers, safety staff, 0 The regulations systems, including any and reactor operators in specialized courses in main safety alarm action positions. sodium technology and LMFBR operator training. Sodium technology courses, such as the pro- Training on the LMFBR simulator should include gram at Cadarache, can be an excellent way to teach the following: the fundamentals in understanding and working with a sodium system; sodium system design and Core incidents, such as the rising or lower- application; and the operation, maintenance, and ing of the control rod safety of sodium systems.

52 COMPARISON OF THE LMFBR HEALTH PHYSICS PROGRAM WITH THE REQUIREMENTS OF THE USNRC STANDARD REVIEW PLAN

The Standard Review Plan for LWRs, NUREG- Regulatory Guide 1.33, Quality Assurali ce Pro- 0800, provides guidance for the USNRC Office of gram Requirements (Operational), will also need Nuclear Reactor Regulation staff responsible for some modification for implementation in review of applications to construct and operate LMFBRs. With the Regulatory Guide modifica- nuclear power plants. The SRP defines the areas of tion, the supporting NUREGs and ANSI Stand- review, acceptance criteria, review procedures, eval- ards that address the quality requirements for uation findings, and implementation requirements PWRs and BWRs will need to be modified hecause for the applicant’s safety analysis report. of the sodium circuits in LMFBRs.

Those parts of the plan pertaining to occupa- Regulatory Guide 8.8, Information Relevant to tional health physics were reviewed in this study to Insuring that Occupational Radiation Exposures at evaluate the Plan’s requirements in light of experi- Nuclear Power Plants will be As Low As ReaAonably ence gained at operating LMFBRs. The following Achievable, will need very little modification for sections of the SRP were reviewed Section 12.1, implementing reference to LMFBRs. For example, which discusses how the SAR is reviewed as it Section 2, “Facility and Equipment Design Fea- relates to ensuring that occupational radiation tures,” will need additional information on fission exposures (ORE) will be as low as is reasonably product source term estimates for LMFBRs. Sec- achievable; Section 12.2, which relates to radiation tion 2E, “Crud Control,” will need guidance sources in normal operations, anticipated opera- applicable to LMFBRs. Section 2F, “Isolation and tional occurrences, and accident conditions affect- Decontamination,” will require guidance on the ing in-plant radiation protection; Sections 12.3 decontamination of sodium wetted materials and and 12.4, which relate to radiation protection equipment. Section 2H, “Resin and Sludge Treat- design features described in the SAR, taking into ment Systems, ” has very little applicability to account design dose rates, anticipated operational LMFBRs, as it mostly applies to water circuits. It occurrences, and accident conditions; and would be desirable to have information in this sec- Section 12.5, which relates to the operational radi- tion on guidance for handling contaminated ation protection program. sodium. Sections 3 and 4 of Regulatory Guide 8.8 will not require extensive modifications for its The acceptance criteria for Section 12.1 are implementation in LMFBRs. However, Section 4D included in various subsections of 10 CFR 19 and on “Protective Equipment” should providc guid- 20, and several regulatory guides and NUREGs ance for the protective clothing, equipment, and that provide information, recommendations, and respiratory protective equipment for work ing in guidance, and in general describe a basis acceptable sodium aerosols or around sodium fires. to the NRC staff for implementing the requirements of 10 CFR 19 and 20. Section 12.2 of the SRP describes the requirements in the SAR as they relate to radiation sources The requirements for notices, instructions, and in normal operations, anticipated operational reports by licensees to individuals participating in occurrences, and accident conditions. It requires in license activities are established in 10 CFR 19. the SAR a description of contained sources, air- Standards for protection against radiation are dis- borne radioactive material sources, and the accept- cussed in 10 CFR 20. The REG Guides and ance criteria that must be met in controlling these NUREGs used for implementing the requirements sources. Additional recommendations and guid- of 10 CFR 19 and 20 will require modification for ance will be needed for the licensing of LMFBRs in the licensing of LMFBRs. this area. Specific guidance will be required for evaluating the potential radiological consequences Regulatory Guide 1.8, Personnel Selection and of loss of fluid in core disruptive accidents in Training, used as a basis to the NRC staff for com- LMFBRs. Sufficient operating experience has been plying with the commission’s regulations with gained in the test, experimental, and demonstra- regard to qualifications of radiation protection per- tion LMFBRs to provide the typical long-term con- sonnel, will need some modification to entail centrations of principal radionuclides In the requirements for the training of personnel in various circuits of the LMFBR power plants and sodium technology. the release of radioactive materials from LMFBRs.

53 The basis used in developing shielding and ventila- Standard technical specifications for LMFBRs tion design fission product source terms in the SRP as they relate to radiation protection considerations n will have to be modified to include LMFBRs. and the applicability, format, and implementation of the technical specification package need to be Specific REG Guides and NUREGs need to be developed. The ANSI standards and NUREGs developed for providing information, recommen- listed in the SRP as recommendations, guidances, dations, and guidance concerning the assumptions or information for LMFBR licensing should be to be used for evaluating radiological consequences acceptable for implementing the requirements of of postulated accidents at LMFBRs. Examples of Sections 12.3-12.4 with respect to sampling air- accidents at LMFBRs to be considered are steam borne radioactivity, location and design criteria for generator leaks, Na/water reactions, fuel failure area radiation monitors, criticality accident alarm propagation, rupture of primary pumping, pump system, concrete radiation shields, and general failure or reactivity transient with the Plant Protec- reactor shielding at LMFBRs. tion System operating, and successive failures of normally provided and maintained multiple barri- Standard technical specifications for LMFBR ers. Specific acceptance criteria for LMFBRs are designs as they relate to radiation protection con- necessary as required for LWRs in Regulatory siderations in the applicability, format, and imple- Guide 1.112 and ANSI N237-1976 with regard to mentation of the LMFBR construction firms radiological source terms. As was done for LWRs in technical specification package will need to be writ- NUREG-07 18 and 0737, the basis for acceptable ten, as they were for LWRs in NUREG-0103, -0123, shielding and ventilation design fission product -0212, and -0718. source terms needs to be developed.

In Sections 12.3-12.4, “Radiation Protection The administrative organization of the radia- Design Features in the SRP,” the radiation zone tion protection program (described in SRP designations during accident conditions are Section 12.5), including the experience and qualifi- defined based on Regulatory Guides 1.3, 1.4, and cation of the personnel and the information 1.7. Regulatory positions need to be developed to described for the implementation of Regulatory provide future licensees with assumptions to be Guides 1.8, 8.2, 8.8, and 8.10, will not be signifi- used for evaluating the potential radical conse- cantly different for LMFBRs. The description of quences of loss-of-coolant accidents or other high- the bases and methods for monitoring and control consequence accidents applicable to LMFBRs. of surface contamination for personnel and equip- Table 19 describes accident conditions categorized ment, including the reporting practices for normal by ANS Standards. and off-normal or accident conditions, may need to be modified. Special air sampling and the issu- Ventilation requirements in Sections 12.3-12.4 ance and use of respiratory equipment will be dif- of the SRP will need to be modified with regard to ferent at LMFBRs because of the effects of their component design criteria and qualification sodium. The description of physical and adminis- testing to ensure compatibility with sodium sys- trative measures for controlling access and stay tems. Additional criteria for component design and times in radiation areas, the procedures and meth- qualification testing need to be developed for the ods of operation for ensuring that occupational engineered safety features of the atmospheric radiation exposure will be as low as reasonably cleanup system air filtration of LMFBRs. Criteria achievable; and the description of methods, fre- need to be developed for instrumentation for quency, and procedures for conducting radiation LMFBRs to assess plant conditions during and fol- surveys will not be significantly different at lowing an accident. The criteria should include the LMFBRs. Personnel protection equipment may following: have to be modified to be applicable to LMFBRs, because of the potential sodium spill and fires. Variables to be monitored Facilities and equipment to clean, sanitize, repair, and decontaminate personal protective equipment, Design basis accident events monitoring equipment respirators, and other The duration various instruments are qual- equipment involved in radiation protection, may ified to function need to be different at LMFBRs. These differences again are based on the use of sodium instead of The variable ranges for monitoring. water in the cooling system.

54 Table 19. Accident conditions categorized by ANS standards

Category Examples

PWR and BWRa

I. Normal Operation Startup, shutdown, standby, power ascension from partial load to full power, cladding defects within Expected frequently during normal course of oper- tech specs, refueling ation and maintenance

11. Incidents of Moderate Frequency Inadvertent control rod group withdrawal, partial loss of core cooling, moderate cool-down, loss of Any one may occur during a calendar year for a off-site power, single error of operator particular plant

111. Infrequent Events LDSS of reactor coolant (with normal coolant makeup system only), secondary pipe break, fuel May occur during the lifetime of a particular plant assembly in violation of tech specs, control rod withdrawal in violation of tech specs, unexplained reactivity insertion, complete loss of core flaw (excluding pump locked rotor)

IV. Limiting Faults Major pipe rupture (up to and including double- ended rupture of largest pipe), fuel or structure Not expected to occur, but postulated because of movement due to core damage, ejection of single potential for significant radioactive release; most control rod, major secondary system pipe rupture drastic faults that must be designed against (double-ended), coolant pump locked rotor

I. Normal Operation Startup, normal shutdown, standby, load following, cladding defect with tech specs, refueling Expected frequently during normal course of operation and maintenance

11. Anticipated Operational Occurrences Tripping of Na pumps, failure of all off-site power, tripping of TurbineGenerator set, inadvertent con- May individually occur one or more times during trol rod withdrawal the lifetime of the plant

111. Postulated Accidents Spectrum of events appropriate to a specific design considering both the probability and conseqiiences Not expected to occur, but included in the design of the events (e.g., pipe rupture, large Na fire, large basis to provide additional margins to ensure no Na-water reactor, rupture of radwaste system tank) undue risk to the health and safety of the public a. For a PWR, see Reference 187.

For a BWR, see Reference 188. b. For an LMFBR, see References 189 and 190.

55 FUTURE RESEARCH REQUIREMENTS n

Further research and development are still neces- Studies external to containment are required for a sary in the field of occupational health physics. better understanding of the radiological source term Developmental needs related to LMFBRs are for accidents and off-normal operations at LMFBRs. health physics instrumentation, secondary contain- The present knowledge on sodium aerosols has ment studies, studies external to containment, and shown that their inherent attenuation capability could probabilistic risk analysis. be very important within containment. Trapping of gaseous and volatile fission products by the aerosols and high aerosol deposition studies have been shown Health physics instrumentation and personnel to be significant. The engineered attenuation systems dosimetry techniques will need to be perfected for are very efficient in reducing the overall radioactivity use in the extremely high beta-gamma radiation release through the containment. Codes used in mod- fields that can be encountered during maintenance eling the transfer of aerosols to the atmosphere need on LMFBR components. During repair and main- to be adapted to LMFBRs. The consequences of tenance on sodium pumps at EBR-11, the beta- plume depletion, ground deposition, resuspension, gamma dose rates in air were found to be 10,000: 1. and sodium aerosol chemistry, and the kinetics of The source for this high beta dose rates was Sr/Y-90 transformation to flora and fauna and personnel dose that was impinged upon the impeller of the pump. must be studied. The Sr/Y-90 was not removed during the normal Probabilistic risk assessment for optimization of steam/alcohol/water cleaning process. The typical radiation exposures to workers and the public needs health physics instrumentation used in surveying to be developed for LMFBRs. Analytical tech- the pump components was not able to measure the niques are needed for quantifying the implementa- intensity of the beta fields. Improved technology tion of ALARA programs so that proper and applied techniques, as well as clear guidance, cost-effectiveness criteria can be defined. Such will be necessary to provide the worker with suffi- techniques are presently being developed for opti- c ient protection. Current dosimeters, instruments, mizing the balance of worker dose versus increment and field control techniques for detecting, evaluat- reductions of public exposure and costs. These ing, and measuring nonpenetrating dose in high techniques are needed for LMFBRs. beta fields are deficient. The effects of sodium chemistry on the physio- logical behavior of activation and fission products Secondary containment studies with regard to should be studied. Recommendations by the sodium leaks and fires will require study. Develop- National Council on Radiation Protection and ment is needed of ventilation systems and filter sys- Measurement (NCRP) and the International Com- tems for sodium aerosols produced during fires. mission on Radiological Protection assume certain Prefilters like aqueous scrubbers and the perform- physiochemical states for many radionuclides as a ance of HEPA filters need to be studied. Sensors basis for estimating the doses to tissues and organs. with faster response than the presently used hydro- There may be significant differences in these states gen detector for small sodium leaks, less than for particular radionuclides because of the sodium 50 gal/% are needed. coolant at an LMFBR.

56 REFERENCES i 1. L. P. Abagyan et al., Group Constant for Nuclear Reactor Calculations, Consultants Bureau Enter- prises, Inc., New York, New York, 1964.

2. E. Storm and H. I. Israel, Photon Cross Sections from 0.001 to 100 MeVforElements 1 throlrgh 100, Los Alamos National Laboratory, LA-3753, 1967.

3. The ORNL Benchmark Experiment for Neutron Transport in Thick Sodium, Oak Ridge National Laboratory, ORNLTM-3969.

4. Final Report on a Benchmark Experiment for Neutron Transport through Iron and Stainless Steel, Oak Ridge National Laboratory, ORNL-4892, 1973.

5. R. Rittenberger, J. Gibson, and J. Ritts, Radial Blanket and Shield Simulation Experiments-A nalysis of Experimental Reports, Topical Report, Westinghouse, WARD-SS-3045-3, 1973.

6. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Labclratory, ORNL-TM-4364, 1973.

7. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Laboratory, ORNL-TM-4420, 1973.

8. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Laboratory, ORNLTM-4469, October 1973.

9. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Laboratory, ORNL-TM-45 14, November/December 1973.

10. FAST Reactor Experimental shielding Monthly Progress Report, Oak Ridge National Laboratory, ORNL-TM-4592, January 1974.

11. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Lab0 ratory, ORNL-TM-4611, February 1974.

12. FAST Reactor Experimental Shielding Monthly Progress Report, Oak Ridge National Lab0 ratory, ORNL-TM-4402, July/AUgust 1973.

13. FAST Reactor Analytical Shielding Monthly Progress Report, Oak Ridge National Lab0 ratory, ORNL-TM-4482, July/August 1973.

14. W. L. Kuhn, A Computational Techniquefor Estimating Corrosion Product Releases in an LMFBR, Hanford Engineering Development Laboratory, HEDL-SA-1077, 1976.

15. A. Merkel and G. Menken, “Estimation of the Radiation Levels Near the Components in the SNR-300 Primary System Induced by Deposition of Corrosion Products, ” Proceedings of the BNES Interniztional Conference on Liquid Alkali Metals, Nottingham, UK, 1973.

16. T. Yanagisawa, “Design Considerations of Fission and Corrosion Products in Primary System of MONJU,” Paper No. 11, IAEA Specialists Meeting on Fission and Corrosion Products in Primary Systems of LMFBRs, Dimitrovgrad, USSR, 1975.

17. W. E Brehm, “Radioactive Corrosion Product Transport and Control,’’ Proceedings of the Znterna- tional Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976. /\

57 18. R. Mathews, “Location and Repair of a Leak in the Dounreay FAST Reactor Primary Circuit,” n J; British Nuclear Energy Society, 1969.

. 19. A. E. Leypunski, “Study of the Five-Year Operation of the BR-5 Reactor,” Paper No. 312, Third Geneva Conference on Peaceful Uses of Atomic Energy, Geneva, Switzerland, 1964.

20. I. A. Efimov, “BR-5 Primary Circuit Decontamination,” Paper No. 17, IAEA Specialists Meeting, Demitrovgrad, USSR, 1975.

21. V. D. Kizin, “Studies of Radioactivity Buildup and Distribution in the BOR-60 Circuit,” Paper No. 15, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975.

22. H. H. Stamm ,and K. C. Stade, “Corrosion Product Behavior in the Primary Circuit of the KNK Nuclear Reactor Facility, ” Paper No. 5, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975.

23. R. DeFremont, “Decontamination Experience on Rapsodie, ” IAEA Specialists Meeting on Fission and Corrosion Products in Primary Systems of LMFBRs, Bensburg, Germany, in USERDA Report CONF-7 10959, 1971.

24. L. Costa, “Corrosion and Activity in the Primary Circuit of a Sodium Reactor-Code CORONA,” Paper No. 3, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975.

25. W. F. Brehm, “Effect of Oxygen in Sodium upon Radionuclide Release from Austenitic Stainless Steel, ” HEDGSA-985, Paper No. 29, IAEA Specialists Meeting, Dimitrovgrad, USSR, 197.5.

26. J. R. Humphreys, “Problem Considerations in the U.S. Radioactivity Control Technology Pro- gram,” Paper No. 20, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975.

27. T. J. Kabele, Activated Corrosion Product Transport Radiation Level Near FFTF Reactor and Closed Loop Primary System Component, Hanford Engineering Development Laboratory, HEDL-TME-72- 71, 1972.

28. J. N. Anno and J. A. Walowit, “Analysis of Corrosion of Stainless Steel in a Sodium and High Radiation Environment,” Nuclear Echnology, 10, 1971, pp. 67ff.

29. W. F. Brehm, “Radioactive Material Transport and Flowing Sodium System,” in Corrosion by Liq- uid Metals, Plenum Press, New York, New York, 1970.

30. P. Roy, USERDA Report GEAP-10394, 1971.

31. R. P. Colburn, “Transport of Activated Corrosion Product in Sodium Loops,” Transactions of the American Nuclear Society, 17, 1973, pp. 165ff.

32. G. Menken and H. Reichel, “Corrosion and Deposition Behavior of Co-60 and Mn-54 in the SNR Mockup Loop,” Paper No. 6, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1979.

33. N. Sekiguchi, “Behavior of Corrosion Products from Irradiated Stainless Steel and Flowing Sodium,” Paper No. 8, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975.

34. K. T. Claxton and J. G. Collier, “Effects of Corrosion Product Transport in FAST Reactors,” BNES Conference, Nottingham, England, 1973.

35. W. F. Brehm, “Techniques for Studying Corrosion and Deposition of Radioactive Materials in Sodium Loops, ” HEDL-SA-986, Paper No. 18, IAEA Specialists Meeting, Dimitrovgrad, USSR, 1975. n

58 36. B. D. Kizin et al., “Radioactivity in the Primary Circuit of the BOR-60 Fast Reactor,” AT Ehnerg. 44, 1978, pp. 492ff.

37. C. A. Erdman and A. B. Reynolds, “Radionuclide Behavior During Normal Operation of Liquid Metal Cooled Fast Breeder Reactors, Part 1, Production,” Nuclear Safety, 16, 1975, pp. 43ff.

38. N. Z. Krasnoyarov, V. I. Polyakov, and Y. V. Chechetkin, “Radioactive Product Behavior :During the Operation of Sodium Cooled Reactors of the USSR,” Proceedings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, May 3-6, 1976.

39. B. D. Kizin et al., Deposition of Radionuclides by Cold napping in the BOR-60, Kernforschungszentrum, Karlsruhe, KFK-TR-575, NIIAR-P-2‘(296), 1977.

40. C. G. Allan, and T. A. H. Taylor, “Activity Deposition in DFR Primary Cold Trap,” Proceedings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, May 3-6, 1976.

41. C. R. F. Smith and J. T. Holmes, Purity of EBR-11 Sodium, 1967-1974, ANL-75-26, Chicago, Illinois, 1975.

42. J. Krenser and A. Lacroix, “Some Aspects of Sodium Technology Issues from the Operating Experi- ence of Rapsodie and PHENIX,” Proceedings of the International Conference on Liquid Metd Tech- nology and Energy Production, Champion, Pennsylvania, 1976.

43. H. H. Stamm, “Sorption of Radionuclides on Sodium Sampling Vessels,” Proceedings of the Inter- national Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976.

44. M. Boutrais, “Sodium Analysis and Purification: Aerosols Trapping in PHENIX Reactor,” Pro- ceedings of the BNES International Conference on the Fast Reactor Power Station, London, UK, 1974.

45. D. F. Washburn et al., “Mixed Oxide Run Beyond Cladding Breach Test at EBR-11,” Proceedings of the International Conference on Fast Breeder Reactor Fuel Performance, Monterey, California, 1979.

46. J. I. Sackett et al., “Preparations at EBR-I1 for Breached Element Operation,” Transactions of the American Nuclear Societx 27, 1977, pp. 823ff.

47. J. D. B. Lambert, E S. Kim, and E. R. Ebersole, “Experience with Breached Elements at EBR-11,” Transactions of the American Nuclear Society, 27, 1977, pp. 820ff.

48. L. R. Monson et al., “Fission Product Control at EBR-11,” Transactions of the American Nuclear Society, 27, 1977, pp. 822ff.

49. R. R. Matthews and K. J. Henry, ‘‘Location and Repair of a Leak in the Dounreay Fast Reactor Primary Circuit,” J. Br. Nucl. Energy Soc. 8, 1969, pp. 168ff.

50. M. J. Adamson, Sodium-Fuel Reaction Studies: Fuel Pellet Swelling Behavior and Alkali Metal-Oxide Fuel Reaction Thermodynamics, General Electric Company, GEAP- 14093, 1976.

51. P. E. Bohaboy et al., “Performance of Failed Fuel in Fast Reactors,” Proceedings of the biterna- tional Conference on Fast Reactor Power Stations, London, 1974, pp. 349ff.

52. A. Tibi, J. Misraki, and J. P. Broutin, “La Contamination Alpha A Rapsodie,” Fission and Corro- sion Product Behavior in Primary Circuits of LMFBR ’S, Proc. Meeting Dimitrovgrad, USSR, 1975, IAEA Report IWGFR/7, Vienna 12, 1976.

59 53. W. Glauner et al., “Das Defekstavprogramm in SILOE, 1978,” ReaklorErgebnkse Des Ersten I/ersuches, KFK Nacher: 2,21.

54. S. Jacobi and G. Schmitl, “Release of Fission Products From Artificially Defected MFBR Oxide Fuel Pins,” Proceedings of the International Conference on FAST Breeder Reactor Fuel Performance, Monterey, 1979, pp. 607ff.

55. R. M. Fryer, L. A. Reyes, and R. V. Strain, “Response of EBR-I1 Delayed Neutron Monitor to Tramp Uranium and a Fission Product Source,” Trans. Am. Nucl. SOC.,28, 1978, pp. 710ff.

56. A. E. Waltar and A. B. Reynolds, Fast Breeder Reactors, Pergamon Press, Elmsford, New York, 1981.

57. A. Cafasso et al., “The Solubilities of the Noble Gases in Liquid Sodium,” Proceedings of the Interna- tional Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 619ff.

58. R. H. S. Winterton, “Cover-Gas Bubbles in Recirculating Sodium Primary Coolant ,” Nuclear Engi- neering Design, 22, 1972, pp. 262ff.

59. J. Saroul and R. Michael, Transfer of Rare Gases Through Liquid Sodium, Experiments ATENA, Commissariat A 1’Energie Atomique, CEA-R-4579, 1974.

60. N. Aubert and M. Merle, “Gas Entrainment Problems in Sodium Cooled Nuclear Reactors,” Pmceed- ings of the Symposium on Two-Phase Flow and Cavitation in Power Generating Systems, Grenoble, 1976, pp. 293ff.

61. N. V. Polley, Bubbling from LMFBR Fuel Failures and Its Relevance to Failed Fuel Location, Central Electric Generating Board, Harwell, England, CEGB-RD/B/N 2837, 1973.

62. R. W. Kessie, Inert Gas Sodium Contracters, Argonne National Laboratory, ANL-7817, 1971, pp. 44ff.

63. N. R. Chellew et al., “A Method for Detecting Fuel Cladding Penetration in LMFBR’s by Monitor- ing the Sodium Coolant for 1-135,” Nucl. Technol, 21, pp. 125ff.

64. N. Mitsutsuka et al., “Fission Product Measurement in the In-Pile Fission Product Loop,” Proceed- ings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 284ff.

65. G. S. Brunson, “Monitoring Fission Gas in EBR-I1 by High Resolution Gamma Spectrometry,” Nucl. Technol, 25, 1975, pp. 553ff.

66. N. Mitsutsuka, “Method for Determining Degassing Constant of Xenon in Liquid Sodium,” J Nucl. Sci. Technol., 15, 1978, pp. 678ff.

67. W. B. Brenner and A. €3. G. Washington, “Measurement of Fission Gas Release Rates From a U02 Fuel Pin During Irradiation in a Fast Reactor,” J Br. Nucl. Energy SOC., 12, 1973, pp. 449ff.

68. J. R. Findlay, “The Migration of Fission Products Through Reactor Fuel Materials,” Proceedings of Panel on Behavior and Chemical State of Irradiated Ceramic Fuels, IAEA, Vienna, 1974, pp. 21 Iff.

69. N. V. Krasnoyarov, “Release of Gaseous Fission Products From Defect Fuel Pins of the Reactor BOR-60,”At. Ehnerg., 38, 1975, pp. 67ff. ~-

60 70. B. R. Sehgal and R. H. Rempert, Trans. Am. Nucl. SOC.,14, 1971, pp. 779ff.

71. R. J. Budnitz, “Tritium Instrumentation for Environmental and Occupational Monitor ing-A Review,” Health Physics Journal, 26, 1974, pp. 165ff.

72. J. T. Holmes and M. M. Wrightson, Plan for Studying Tritium Transport, Unpublished Argonne National Laboratory Report, 1975.

73. J. T. Bell and J. D. Redman, “Tritium Permeation Through Metals Under Steam Conditions,” Pro- ceedings of the Ninth Symposium on Fusion Technology, Commission of the European Comniunica- tions, Pergamon Press, Oxford and New York, 1976, pp. 345ff.

74. R. A. Strehlow and H. C. Savage, “The Permeation of Hydrogen Isotopes Through Structured Metals at Low Pressures and Through Metals With Oxide Film Barriers,” Nucl. Technol., 22, 1974, pp. 127ff.

75. R. A. Strehlow and H. C. Savage, “Formation and Use of Oxide Films to Impede Outgassing of Metals,” 1 Nucl. Master, 53, 1974, pp. 323ff.

76. J. T. Bell, J. D. Redman, and F, J. Smith, “Oxides as Barriers to Tritium Permeation in Stearn Gen- erators and Tritium Content in Controlled Thermonuclear Reactor Coolants, ” Proceeding of the 24th Conference on Remote Systems Echnologj 1976, pp. 3ff.

77. R. Kumar, Tritium Transport in an LMFBR, Argonne National Laboratory, ANL-80/89, 1974.

78. T. J. Kabele, Parametric Study of the Tritium Distribution in the FFTF, Hanford Engineering Devel- opment Laboratory, HEDL-TME 1974, pp. 74-6.

79. T. A. Renner, D. J. Raue, and C. C. McPheeters, “Tritium Permeation Through Steam Generator Materials,” Trans. Amer. Nucl. SOC.,126, 1977, pp. 216ff.

80. V. A. Maroni et al., “Defects of Surface Environment and Bulk Composition on the Permealion of Hydrogen Isotopes Through Iron and Nickel Based Alloys,” Symposium on Tritium in Nuclem Sys- tems, 107th AIME Annual Meeting, Denver, Colorado, February 26-March 2, 1978.

81. J. C. McGuire, “Tritium Permeation Through Oxidized 304 and 316 Stainless Steel,” Symposium on Tritium in Nuclear Systems, 107th AIME Annual Meeting, Denver, Colorado, February 26-March 2.

82. T. A. Renner and C. C. McPheeters, Tritium and Hydrogen Transport in LMFBR Systems: EBR-11, CRBR, and FFTF, Argonne National Laboratory, ANL-78-64, 1978.

83. Project Management Corporation, Clinch River Breeder Reactor Project Preliminary Safety A.rzalysis Report, vol. 11, see. 11.1.4, Amendment 67, March 1982.

84. H. I. McCreery, A Study of Tritium Activity in the EBR-11 Reactor Building Air, University of Idaho Master’s Thesis, 1977.

85. D. J. Gorman, An Assessment of the Relative Hazards of T2, and T20, Ontario Hydro Report, HPD- 78-1, Toronto, Canada, 1978.

86. L. C. Witbeck and R. M. Fryer, Waste Production and Management at EBR-II, Argonne National Laboratory Report 79-14, 1979.

87. E. R. Ebersole, Tritium Control Technology, WASH 1269, 1973, pp. 286ff.

61 88. Dr. Graham Tyler, Private Communication, Dounreay Nuclear Power Development Association, June 29, 1983.

89. C. S. Abrams, Private Communication, Argonne National Laboratory-West, May 11 , /983.

90. C. C. McPheeters, S. D. Skladzien, and M. M. Osterhout, “Development and Testing of a Tritium Meter for Sodium Systems,” Trans. Amer: Nucl. SOC.,28, 1977, pp, 787ff.

91. S. B. Sklabzien et al., Tritium Meters for Use in LMFBR Sodium Coolant, Argonne National Lubora- tory, ANL78-30, 1978.

92. B. A. Maroni et al., “Solution Behavior of Hydrogen Isotopes and Other Non-Metallic Elements in Liquid Lithium,” Proceedings of the International Conference on Liquid Metal Technology and Energy Production, Champion, Pennsylvania, May 3-6, 1976, pp. 437ff.

’ 93. P. Coulon, Private Communication, PHENIX, July 5, 1983.

94. Graham Tyler, Private Communication, Central Station, PFR, June 29, 1983.

95. K. C. Stade, Private Communication, KNK2, July 8, 1983.

96. P. R. Prevo and G. D. Carpenter, private communication, Hanford Engineering Development Labo- ratory, May 17, 1983.

97. A. H. Kibbey and H. W. Godbee, Critical Review of Solid Radioactive Waste Practices at Nuclear Power Plants, Oak Ridge National Laboratory, ORNL4924, 1974.

98. W. A. Ross, “Characteristics of High Level Radioactive Waste Gasses,” Trans. Amer. Nucl. SOC.,28, 1978, pp. 357ff.

99. J. G. Asquith et al., “Sodium Removal Technology-The Alcohol Process,” Proceedings of the Inter- national Conference on Liquid Metal Technology and Energy Production, Champion, Pennsylvania, May3-6, 1976, 1976, pp. 548ff.

100. H. Richard et al., “Experience From Purification of S0diu.m Wetted Components at the KNK Facil- ity,” Proceedings of the International Conference on Liquid Metal Technology and Energy Production, Champion, Pennsylvania, May 3-6, 1976, 1976, pp. 561ff.

101. C. B. Johnson (ed.), “Sodium (NaK) Supplement, ” Liquid Metals Handbook, TID-5277.

102. C. W. Funk, “Review of Sodium Removal Processes Using Water,” Proceedings of the International Conference on Liquid Metal Technology and Energy Production, Champion, Pennsylvania, May 3-6, 1976, 1976, pp. 554ff.

103. J. C. Commander et al., Decontamination and Decommissioning of the EBR-I Complex, Topical Report No. 3 NaK Disposal Pilot Test, Aerojet Nuclear Company, ANCR-11700, 1975.

104. E. W. Kendall and C. K. Wang, Decontamination and Decommissioning of the EBR-I Complex, Final Report, Aerojet Nuclear Company, ANCR-1242, 1975.

105. M. Sittig, Sodium: Its Manufacture, Properties, and Uses, Reinhold Publishing Corporation, New York, 1956, pp. 149ff.

106. DuPont Sodium Handling and Properties, E. I. duPont DeNemours and Company, Inc., Wilmington, Delaware.

62 - ...... -...... - .. ...

107. T. R. Johnson, Method for the Safe Disposal of Alkali Metals, U.S. Patent No. 4,320,615, .lune 28, 1977.

108. C. C. McPheeters, and R. D. Wolson, Disposal of Radioactive Sodium Waste, Argonne National Laboratory, ANL-78-87, 1978.

109. L. Whitback, Private Communication, Argonne National Laboratory-West, May 10, 1983.

110. K. C. Gross, Fission Product Releases to the Primary System of EBR-IIfrom May 1979 to March 1980, Argonne National Laboratory, ANL-81-64, 1981.

11 1. R. R. Smith et al., LocatingandIdentifyingtheSourceof theMay24, 1967, Fission Product Release to EBR-II, Argonne National Laboratory, ANL-75-43, 1969.

112. R. R. Smith, Origin of Fission Product Releases in EBR-11, November 23, 1967 to May 6, 1968, Argonne National Laboratory, ANL-7604, 1970.

113. R. M. Fryer et al., Symptoms and Detection of a Fission Product Release from an EBR-II Fuel Ele- ment: Case I, Defect Above Fuel Elevation, Argonne National Laboratory, ANL-7605, 1970.

114. R. M. Fryer et al., Symptoms and Detection of a Fission Product Release from an EBR-II Eye1 Ele- ment: Case 2, Defect Below Fuel Elevation, Argonne National Laboratory, ANL-7676, 1970.

115. G. S. Brunson, R. M. Fryer, and R. V. Strain, Experimental Irradiations of Fuel Elements Having Known Cladding Defects in EBR-II, Argonne National Laboratory, ANL-7782, 197 1.

116. R. R. Smith et al., Fission Product Releases to the Primary System of EBR-IIDuring 1971, Argonne National Laboratory, ANL-7956, 1972.

117. D. Y. C. So et al., Fission Product Releases to the Primary System of EBR-II from January 1974 to March 1975, Argonne National Laboratory, ANL-76-65, 1976.

1 18. B. Y. C. So et al., Fission Product Releases to the Primary System of EBR-II from April 1975 to March 1977, Argonne National Laboratory, ANL-79-26, 1979.

119. J. D. B. Lambert et al., “Recent Improvements in Identifying Fission Product Sources in the Experi- mental Breeder Reactor-11,” Nucf. Technol, 39, 1978, pp. 275ff.

120. B. Y. C. So et al., Fission Product Releases to the Primary System of EBR-II from April 1977 foMay 1978, Argonne National Laboratory, ANL79-27, 1979.

121. B. Y. C. So et al., Fission Product Releases to the Primary System of EBR-11 from June 1978 to April 1979, Argonne National Laboratory, ANL-80-102.

122. G. S. Brunson, “Monitoring Fission Gas in EBR-I1 by High Resolution Gamma Spectrometry,” Nucl. Technol. 26, 1975, pp. 553ff.

123. G. S. Brunson, Germanium-Lithium Argon Scanning System (GLASS): Design and Experience Through 1974, Argonne National Laboratory, ANL-76-74, 1976.

124. K. G. Porges, EBR-II Fuel Element Failure Detectors, Reactor Physic.s Division Annual Rqgort- July I, 1963 to June 30, 1964, Argonne National Laboratory, ANL-7010, 1965, pp. 231ff.

125. R. M. Fryer et al., Efficiency of the EBR-II FERD System as an Automatic Trip Device, Argonne National Laboratory, ANL-76-94, 1977.

63 126. K. C. Gross et al., Development of a Methodology for Analysis of Delayed Neutron Signals, Argonne National Laboratory, ANL-79-67, 1980. Q 127. K. C. Gross and R. V. Strain, “Delayed Neutron Siqnal Characterization in a Fast Reactor,” Nucl. Sci. Eng. 76, 1980, pp. 163ff.

128. K. C. Gross et al., Fission Product Releases to the Primary System of EBR-II from April 1980 to March 1981, Argonne National Laboratory, ANL-83-80, 1983.

129. R. Villarreal, Private Communications, Argonne National Laboratory-West, May 11, 1983.

130. D. R. Seidel and R. E. Einziger, “In-Reactor Cladding Breach of EBR-I1 Driver Fuel Elements,” Proceedings of the International Conference on Radiation Effects in Breeder Reactor Structural Materi- als, Scottsdale, Arizona, June 23, 1977, Metallurgical SOC.of AIME, New York, 1977.

131. L. R. Monson et al., “The EBR-I1 Cover Gas Cleanup System,’’ IAEA, Symposium on Design, Construction, and Operating Experience of Demonstration Fast Breeder Reactors, Bologna, April 10- 14, 1978, p. 465.

132. B. Y. C. So and G. W. Stauffer, “MIXTAG-A Code to Identify the Components of a Mixture of Xenon Tags,” Trans. Am. Nucl. SOC.,26, 1977, pp. 522ff.

133. K. C. Gross, POLYFAIL: A Program for Identification of Multiple Fuel Failures with Gas Tagging, Argonne National Laboratory, ANL80-87, 1980.

134. F. Holze, Private Communications, Westinghouse, Hanford, FFTF, May 19, 1983.

135. K. Wirtz, Lectures on Fast Reactors, Kernforschungszentrum, Karlsruhe, 1973.

136. H. Olson, Private Communication, EBR-11, May 10, 1983.

137. 0. J. Foust (ed.), Sodium-NaK Engineering Handbook, Vols. 1 and 5, Gordon and Breach, Science Publishers, Inc., New York, 1972.

138. C. C. McPheeters et al., Chemical Engineering Division, Sodium Technology Program Annual Pro- gress Report, October 1979 September 1979, Argonne National Laboratory, ANL-81-21, 1981.

139. P. Roy and D. N. Rogers, “Hydrogen Burden from the Steam Side Corrosion in Sodium Heated Steam Generators,” Nucl. Technol. 39, 1978, pp. 213ff.

140. C. C. McPheeters et al., Impurity Control Coolant Technology Program Semiannual Progress Report for Period Ending October 31, 1981, Oak Ridge National Laboratory, ORNL/MSP/1.9-82/1, 1981.

141. S. J. Rogers, Topical Report No. 9-Analysis of High Oxygen and High Carbon Cold Traps, MSA Research Corporation, Evans City, Pennsylvania, MSAR 67-203, 1967.

142. J. Herb et al., Examination of the Enrico Fermi Sodium Cold Trap, Westinghouse Electric Corporation, Pittsburg, WCAP-43-21, 1965.

143. M. M. Osterhout, “Control of Oxygen, Hydrogen, and Tritium in Sodium Systems at Experimental Breeder Reactor-11,” Trans. Am. Nucl. SOC.30, 1978, pp. 159ff.

144. C. C. McPheeters and D. J. Raue, Computer Model for Analyzing Sodium Cold Traps, Argonne National Laboratory, ANL-83-36, 1983. n

-. 145. V. M. Kolda and R. D. Wolson, A Design Study of High Performance Cold Traps for Removal of Hydrogen from Liquid Sodium, Argonne National Laboratory, ANL-83-83, 1983.

146. D. J. Hebditch et al., Sodium Purity Controlin CRF-MechanismsandRates of CrystalNucleatron and Growth, Central Electric Generating Board, Harwell, England, R/M/N 775, 1974.

147. D. J. Hebditch and B. J. Gliddon, “Impurity Crystallization in Liquid Sodium Systems,” Proceed- ings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 43ff.

148. J. Kremser, “Discussion, ” Proceedings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 58ff.

149. P. Coulon, Private Communication, PHENIX Nuclear Plant, Avignon, July 5, 1983.

150. M. Berlin and M. Cauvin, “Comporement des Radioiodides Dans le Circuit Primaire D’un Rcacteur Rapide, ” Fission and Corrosion Product Behavior in Primary Circuits of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 6ff.

15 1. R. Passcard, verbal presentation, IAEA Specialists Meeting on Fission and Corrosion Product Behav- ior in Primary Circuits of LMFBRs, Dimitrovgrad, USSR, 1975.

152. V. V. Orlov et al., “Some Problems of Safe Operation of the BR-5 Plant ,” Proceedings of the Interna- tional Conference on Safety of Fast Reactors, Aix-en Provence, 1967, Bull. In$ Sei. Tech., 1968, pp. 75ff.

153. B. I. Polyakov et al., “Efficiency of the Purification of the Sodium Coolant in the BOR-60 Reactor from Radioactive Impurities by Means of a Cold Oxide Trap,” At. Ehnerg. 38, 1975, pp. 171f’f.

154. J. T. Holmes et al., “Sodium Technology at EBR-11,” Proceedings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 12ff.

155. J. T. Holmes et al., “Sodium Purification by Cold Trapping at the EBR-11,” Nucl. Technol. 32: 1977, pp. 304ff.

156. C. C. McPheeters and D. J. Raue, Computer Model for Analyzing Sodium Cold Traps, Argonne National Laboratory, ANL-83-36, 1983.

157. M. H. Cooper and G. R. Taylor, “Cesium Sorption from Liquid Sodium,” Trans. Am. Nucl. Soc. 11, 1968, pp. 526ff.

158. M. H. Cooper and G. R. Taylor, “Cesium Cold Trapping in a Forced-Convection Sodium System,” Trans. Am. Nucl. SOC.12, 1969, pp. 611ff.

159. M. H. Cooper and G. R. Taylor, “Measurement of Cesium Sorption Isotherms in Liquid Sodium,” Nucl. Technol. 12, 1971, pp. 83ff.

160. J. Guon, “Studies of Cesium Deposition in a Sodium/Stainless Steel System,” Trans. Am. Nucl. Soc., 12, 1969, pp. 612ff.

161. J. Guon, “Effect of Surface/Liquid Partition on the Analysis of Impurities in a Sodium System,” Trans. Am. Nucl. Soc. 14, 1971, pp. 625ff.

162. H. Shinojima et al., “Fission Products Deposition in a Stainless Steel Sodium System,” Fission and Corrosion Product Behavior in Primary Circuits of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 96ff. 65 163. N. Mitsutsuka et al., “Fission Product Deposition in a Stainless Steel Sodium System,” Proceedings of the International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 1976, pp. 291ff.

164. N. Mitsutsuka, et al., “Cold Trapping of Fission Products in a Stainless Steel Sodium Loop,” J. Nucl. Sei. Technol. 14, 1977, pp. 135ff.

165. N. Sagawa et al., “Transport and Deposition of Metals in a Sodium Stainless Steel System: 111. Deposition of Cesium in Stainless Steel Capsules under a Large Temperature Gradient,” J. Nucl. Sei. Technol. 12, 1975, pp. 413ff.

166. G. D. Zwetziq et al., “The Distribution of Fission Products in LMFBR Sodium Systems,” Proceed- ings of the International Conference on Sodium Technology in Large Fast Reactor Designs, Argonne National Laboratory, ANL-7520, Part 1, 1968, pp. 527ff.

167. R. P. Colburn, “Fission Product Removal from Sodium by Hydride Slagging,” Trans. Am. Nucl. SOC.15, 1972, pp. 235ff.

168. C. G. Allan, “Solubility and Deposition Behavior of Sodium Bromide and Sodium Iodide in Sodium/Stainless Steel Systems,” Proceedings of the BNESZnternational Conference on LiquidAlkali Metals, Nottingham, 1973, pp. 159ff.

169. N. H. Cooper et al., “Behavior of Iodine in Sodium Systems,” Trans. Am. Nucl. SOC. IS, 1972, p. 232.

170. B. D. Kizin et al., “Studies on Radioactivity Buildup and Distribution in the BOR-60 Circuit,” Fission and Corrosion Product Behavior in Primary Systems of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 150ff.

171. J. C. Clifford et al., “Behavior of Fission Products in Sodium,” ZAEA Symposium on Alkali Metal Coolants, Vienna, 1966, pp. 759ff.

172. G. Menken and H. Reichel, “Corrosion and Deposition Behavior of CO-60 and Mn-54 in the SNR Mockup Loop for the Primary Sodium System,’’ Fission and Corrosion Product Behavior in Primary Systems of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 63ff.

173. N. Sekiguchi et al., “Behavior of Corrosion Product from Irradiated Stainless Steel in Flowing Sodium,” Fission and Corrosion Product Behavior in Primary Systems in LMFBRs, IAEA Report IWGFR/7, 1976, pp. 82ff.

174. W. F. Brehm et al., 1970, “Radioactive Material Transport in Flowing Sodium Systems,” Corrosion by Liquid Metals, Plenum Press, New York, pp. 97ff.

175. W. F. Brehm et al., “Techniques for Studying Corrosion and Deposition of Radioactive Materials in Sodium Loops,” Fission and Corrosion Product Behavior in Primary Circuits of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 172ff.

176. K. T. Claxton and J. G. Collier, “Mass Transport of Stainless Steel Corrosion Products in Flowing Liquid Sodium,” J. Br. Nucl. Energy Soc. 12, 1973, pp. 63ff.

177. H. H. Stamm and K. C. Stade, “Corrosion Product Behavior in the Primary Circuit of the KNK Nuclear Reactor Facility, ” Fission and Corrosion Product Behavior in Primary Circuits of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 36ff.

66 178. L. Costa et al., “Corrosion et activative dans le circuit primaire d’un reactor au sodium-code,” Corona, Fission and Corrosion Product Behavior in Primary Circuits of LMFBRs, IAEA Report IWGFR/7, 1976, pp. 17ff.

179. S. B. Fisher, “Atomic Diffusion in Alloys Exposed to Liquid Sodium,’’ J. Nucl. Mater. 62, 1!>76, pp. 247ff.

180. J. C. McGuire and W. E Brehm, Progress in Radionuclide Trap Development, Hanford Engineering Development Laboratory, HEDL-TME-77-85, 1977.

181. J. C. McGuire and W. F. Brehm, “A Radionuclide Trap for Liquid Metal Cooled Reactors,” Rzns. Am. Nucl. SOC.30, 1978, pp. 189ff.

182. A. J. Hooper, “Application of Physical Examination Techniques in the Study of Corrosion and Dep- osition in Liquid Metals,” Proceedings of the BNES International Conference on Liquid Aikali Metals, Nottingham, 1973, pp. 201ff.

183. E J. Salzano and S. Aronson, “Kinetic Study of the Decomposition of Cesium-Graphite Lamdlar Compounds,” J. Chem. Phys., 42, 1966, pp. 1323ff.

184. W. S. Clough and S. W. Wade, 1971, “Cesium Behavior in Liquid Sodium-The Effect of Carbon,” J. Nucl. Energy. 25, pp. 445ff.

185. A. W Thorley and C. Tyzack, “The Carborization of Stainless Steels in Sodium Containing Carbon Impurities and Its Effect on Mechanical Properties, ” DNES International Conference on the Efects of Environment on Material Properties in Nuclear Systems, London, 1971.

186. L. R. Monson et al., “Fission Product Control at EBR-11,” Trans. Am. Nucl. SOC. 27, 1977, pp. 822ff.

187. Nuclear Safety Criteria for the Design of Stationary Pressurized Water Reactor Plants, American Nuclear Society, La Grange Park, Illinois, ANS-51.1, N18.2, 1973.

188. Nuclear Safety Criteriafor the Design of Stationary Boiling Water Reactor Plants, American Nuclear Society, La Grange Park, Illinois, ANSI/ANS-52.1, 1978.

189. General Safety Design Criteria for an LMFBR Nuclear Power Plant, American Nuclear Society, La Grange Park, Illinois, ANS-54.1, Trial Use Draft, April 1975.

190. LMFBR Safety Classification and Related Requirements, American Nuclear Society, La Grange Park, Illinois, ANS-54.6, Trial Use Draft, October 1979.

67 APPENDIX

THEORY AND DESIGN OF LIQUID METAL FAST BREEDER REACTORS

A- 1

CONTENTS

THEORY AND DESIGN OF LIQUID METAL FAST BREEDER REACTORS ...... A-1

Physicsof Breeding ...... A-9

Major Design Objectives ...... A-11

Mechanical and Thermal Systems Design ...... A-11

Core and Blanket Arrangements ...... A-11 Fuel Lattice ...... A-13 Fuel Assembly ...... A- 14 Vessel Internals ...... a-14 System Heat Transfers ...... A-15 Components ...... A-21 Shielding ...... A-29 Instrumentation ...... A-35 Auxiliary Systems ...... A-41

OPERATIONAL EXPERIENCE AT LMFBRS ...... A-44

Early Experimental LMFBRs ...... A-44

Clementine ...... A-44 EBR-I ...... A-47 LAMPRE-I ...... A-56 SEFOR ...... A-63 USSRBR-5 ...... A-74

Early Power and Test Reactors ...... A-79

Federal Republic of Germany ...... A-80 France ...... A-87 United Kingdom ...... A-92 United States ...... A-99 Japan ...... a-159

Prototype Reactors ...... A-167

PHENIX ...... a-167

REFERENCES ...... a-177

FIGURES

A-1 . Uranium-238 Plutonium-239 conversion chain ...... A- 10

A.2 . Neutrons produced per absorption as a function of energy for fissile isotopes ...... A-IO

A.3 . External and internal breeding configuration ...... a-12

A.4 . Typical homogeneous and heterogeneous LMFBR core/blanket arrangements incorporating internal breeding ...... A-13

A-3 A.5 . Typical LMFBR fuel assembly system ...... A-15 n A.6 . Typical loop-type LMFBR vessel internals ...... A-16

A.7 . LMFBR heat transport system ...... A-17

A.8 . Plain view of Super Phenix pool. showing the four loops. four primary pumps. and eight intermediate heat exchangers ...... A-20

A.9 . LMFBR steam cycles ...... A-21

A.10 . The Clinch River Breeder Reactor project design for plant control ...... A-22

A-1 1 . Reactor vessel for the loop system of SNR-300 ...... A-23

A.12. Reactor tank for the pool system of Super Phenix ...... A-24

A.13 . Details of the Super Phenix reactor tank and internals ...... A-25

A.14 . Super Phenix’s intermediate heat exchanger penetration into an insulated internal tank . . A-26

A.15 . Construction of Super Phenix’s roof/shield deck ...... A-26

A.14 . Intermediate heat exchanger designs for the Super Phenix pool system and for the CRBRPloopsystem ...... A-28

A.17 . Temperature distributions in an integral superheater ...... A-29

A.18 . Steam generator designs ...... A-30

A.19 . Instability in rise in wall temperature at the transition from nucleate to film boiling ...... A-32

A.20 . Key shielding areas in CRBRP reactor and reactor enclosure system ...... A-34

A-2 1 . Closure head assembly configuration of CRBRP ...... A-35

A.22 . Flux monitoring set for CRBRP design ...... A-37

A.23 . A typical thermowell assembly incorporating a resistance temperature detector (RDT) for measuring sodium temperatures ...... A-38

A.24 . Magnetic flowmeter and flow tube for large sodium coolant piping ...... A-38

A.25 . Simplified diagram of eddy current flowmeter ...... A-38

A.26 . Pressure sensor installation ...... A-39

A.27 . FFTF cover gas monitoring system ...... A-40

A.28 . Induction sodium level probe schematic ...... A-41

A.29 . Trace heating and insulation for 1- to 6.in . pipe ...... A-42 A.30 . A typical sodium cold trap ...... A-43 n

A-4 A-3 1 . Cross section to the CLEMENTIN E reactor shield ...... A-46

. A.32 . EBR-I heat transport system ...... A-47

A.33 . EBR-I Horizontal cross section at midpoint of MARK-I1 core ...... A-48

A.34 . EBR-I Mark-I1 reactor-cutaway view ...... A-49

A.35 . EBR-I Mark-111 fuel rod ...... A-52

A.36 . EBR-I Mark-I11 fuel and blanket assemblies ...... A-52

A.37 . EBR-I Mark-I11 core cross section ...... A-53

A.38 . EBR-I Mark-I11 inner tank assembly ...... A-54

A.39 . LAMPRE-Iplanview ...... A-57

A.40 . LAMPRE-I elevation ...... A-58

A.41 . LAMPRE-I horizontal cross section ...... A-58

A.42 . LAMPRE-I vertical cross section ...... A-59

A.43 . LAMPRE-I fuel capsule and slug ...... A-60

A.44 . LAMPRE-I fuel subassembly ...... A-60

A.45 . LAMPRE-I shield, vertical cross shielding ...... A-61

A.46 . Reactivity loss during LAMPRE-I operation ...... A-62

A.47 . LAMPRE-I core power and inlet and outlet temperature changes from 45 kW(t)vs . time ...... A-62

A.48 . LAMPRE-I transfer functions for various power levels ...... A-62

A.49 . LAMPRE-I plutonium-iron fuel density vs . temperature ...... A-63

A-50 . SEFOR reactor building, elevation view ...... A-64

A-5 1 . SEFOR reactor vessel. elevation ...... A-66

A.52 . SEFOR reactor vessel. plan view ...... A-67

A.53 . SEFOR coolant system ...... A-67

A.54 . SEFOR reactor vessel nozzle elevation ...... A-68

A.55 . SEFOR reflector control system ...... A-70

A.56 . SEFOR reflector control drive ...... A-71

A.57 . General arrangement of the USSR fast reactor BR-5 ...... A-75

A- 5 A.58 . BR-5 fuel assembly ...... A-76

A.59 . Vertical cross section of RAPSODIE reactor ...... A-88

A.60 . Horizontal cross section of RAPSODIE reactor ...... A-89

A.61 . Horizontal cross section of RAPSODIE reactor and shield ...... A-89

A.62 . RAPSODIE fuel subassembly ...... A-90

A.63 . Dounreay plant elevation ...... A-93

A.64 . Details of Dounreay core tube nest ...... A-94

A.65 . Details of Dounreay fuel subassembly ...... A-95

A.66 . Arrangement of Dounreay primary circuit pipe work ...... A-95

A.67 . Cutaway view of the Dounreay reactor vessel and rotating shields ...... A-97

A.68 . EBR-I1 major reactor components and reactor primary tank ...... A-101

A.69 . EBR-IIgridplenumassembly ...... A- 102

A.70 . EBR-I1 reactor vessel assembly ...... A- 102

A.71 . EBR-IIprimaryneutronshield ...... A-103

A.72 . EBR-I1 safety rod drive system ...... A- 104

A.73 . EBR-I1 control and safety rod drive systems ...... A-105

A.74 . EBR-I1 steamgenerator ...... A- 106

A.75 . EBR-I1 principal fuel handling components ...... A- 107

A.76 . Vertical cross section of EBR-I1 reactor building ...... A- 108

A.77 . EBR-I1 blast shield and biological shield ...... A- 109

A.78 . Perspective view of FERMI reactor ...... A-115

A.79 . FERMI plant layout ...... A-116

A.80 . FERMIreactorvessel ...... A-117

A.81 . FERMI meltdown section ...... A-118

A.82 . FERMI.reactor cross section ...... A-118

A.83 . Isometric view of FERMI fuel assembly ...... A-119

A.84 . FERMI fuel subassembly (external side view) ...... A-120

A.85 . FERMI radial blanket subassemblies ...... A-122 n

A-6 ......

A.86 . FERMI reactor vessel arrangement ...... a-i23

A.87 . Vertical cross section of FERMI reactor building ...... A- 124

A.88 . Perspective view of the segment removal operation ...... A-130

A.89 . FFTF reactor cutaway ...... A-136

A.90 . FFTF reactor internals ...... A-137

A.91 . FFTF instrument tree schematic ...... A-138

A.92 . FFTF nuclear control system ...... A-139

A.93 . FFTF core support structure and core barrel ...... A- 140

A.94 . Arrangement of core strain mechanisms around core ...... A-141

A.95 . Location of in-vessel storage modules ...... A-142

A.96 . In-vessel storage module ...... A- 142

A.97 . Low-level flux monitor assembly ...... A- 143

A.98 . Dump heat exchanger module ...... A- 145

A.99 . Sodium flow through FFTF reactor vessel ...... A-146

A.100 . Heat transfer system primary loop piping ...... A- 147

A.101 . Heat transfer system primary pump with guard vessel ...... A-148

A.102 . HTS intermediate heat exchanger and guard vessel ...... A-150

A.103 . HTS secondary loop schematic ...... A-151

A.104 . HTSsecondarypump ...... A-152

A.105 . HTS DHX Air flow diagram ...... A-153

A.106 . Typical plan of internal structures within the containment vessel ...... A-154

A.107 . Typical cross section of internal structures within the containment vessel ...... A-154

A.108 . JOYOcutawayview ...... A- 160

A.109 . Vertical cross section of reactor and core configuration of JOYO ...... A-161

A- 110 . JOYO core diagram ...... A- 162

A-1 11 . Fuel handling, charge/discharge mechanism ...... A-164

A-1 12. Fuel handling and storage facility ...... A- 165

A-1 13. Plant logitudinal section ...... A-168

A-I A.114 . PHENIX reactor block-cutaway view ...... A-168

A.115 . PHENIX fuel assembly ...... A-169

A.116 . Primarypump ...... A-172

A.117 . Expansion tank and secondary pump ...... A-173

A- 1 18 . PHENIX steam generator ...... A-174

A.119 . PHENIX intermediate heat exchanger ...... A-175

TABLES

A.1 . Goals and design objectives for the breeder reactor ...... A-12

A.2 . Comparison of homogeneous and heterogeneous cores [IO00 MW(e) design] ...... A-14

A.3 . Main design parameters for large fast breeder reactors ...... A-18

A.4 . Calculated cover gas activity for the FFTF design basis. 1 Yo pin failure ...... A-33

A.5 . Contaminants monitored by a typical sodium purification system ...... A-43

A.6 . Early experimental LMFBRs ...... A-45

A-I . Reactivity control requirements ...... A-69

A.8 . Early power and test reactors ...... A-81

A.9 . Gamma spectrum six days after sampling ...... A-99

A.10 . Reactor plant control systems ...... A- 144

A.11 . Summary of routine radiation survey data during FFTF Cycle 1 ...... A-157

A.12 . FFTF personnel exposure summary during Cycle 1 operations ...... A-159

A- 8 APPENDIX

THEORY AND DESIGN OF LIQUID METAL FAST BREEDER REACTORS

Physics of Breeding produces the variations of 7 with energy shown in Figure A-2. As a conceptual guide, the criterion for To achieve breeding, a fertile isotope (thorium- this can be expressed as: 232, uranium-234, uranium-238, plutonium-240) must be converted via neutron capture into a fissile 7 = BR + (1 + L) 64-21 isotope (uranium-233, uranium-235, plutonium- 239, plutonium-240). Figure A-1 shows the fertile where conversion chain. The conversion ratio of fissile material produced to fissile material destroyed is L = unproductively lost neulrons, by greater than unity in breeder reactors. It is possible absorption in any material other than for the in-core conversion ratio to be less than one, fissile or fertile material or by leakage while the breeding ratio (BR) for the entire from the reactor reactor-the core plus the region of the reactor containing fertile fuel, known as the blanket-to be BR = Breeding ratio, fissile material pro- greater than unity. The breeding gain, G, is defined duced by the fissile material as G = BR - 1. For a nuclear reactor to have destroyed. G >0, the fissile isotopes must be chosen very carefully for a given neutron energy spectrum. A A high value, greater than two, for 7 is required for high breeding gain can only be obtained with a fast an acceptable breeding ratio. Figure A-2 shows that neutron spectrum. plutonium-239 is the best choice of fissi18: material in a fast breeder. Averaged over a typical U02- The number of neutrons produced per neutron Pu02 fueled LMFBR spectrum, 7 is equal to absorbed during fission, can be calculated from the about 2.45.A-1 Even higher breeder ratios can be following relationship: obtained from fast reactors that use carbide fuels (UC-PuC) or metal fuels. This is because these fuels have greater fissile and fertile material densities and their average neutron energy is higher.

The time required for a particular breeder reactor where to produce enough fissile material in excess of its own fissile inventory to fuel an identical reactor is v = number of neutrons produced per fis- the reactor doubling time, RDT. Doubling time is sion one of the figures of merit used to compare breeder reactor designs, different fast reactor fuels, and CY = capture-to-fission ratio (ac/af). fuel cycle systems involving many breeder reactors. The reactor doubling time may be expressed in The parameters v and CY are measured quantities, terms of the initial fissile inventory in a re,actor, Mo whereas 7 is a derived quantity. For each of the (kg), and the fissile material gained during one primary fissile isotopes, v is fairly constant for year, M (kg/y), where Mg is a time-averaged dif- neutron energies up to about 1 MeV, about 2.5 for ference tetween the fissile inventory at the begin- uranium-233 and uranium-235 and 2.9 for ning of a year and the fissile inventory at the end of plutonium-239, and slowly rises at higher energy. the year. However, CY varies considerably with energy and between isotopes.*-l For uranium-235 and RDT = Mo/M (A-3) plutonium-239, a rises sharply in the intermediate g energy range between 1 eV and 10 KeV and then drops again at high energy; for uranium-233, a An accurate calculation of Mg is difficult. Mg never rises appreciably. This behavior of v and CY can be approximated in terms of the breeding gain,

A- 9 (n,u) h-d O(TlI2 = 14.3 years) 1 0(TlI2 = 4.95 hr) 2%PU

N112 = 12.1 days) 237Np -238Np (",?) IB(TlI2 = 6.75 days)

Figure A-1 . Uranium-238 Plutonium-239 conversion chain.

4.0

3.0

L a, Q 2.0 .-2 a,> c c2 3 1.o za, Note: 17 2 2 + losses for breeding

1 Energy of neutron absorbed

Figure A-2. Neutrons produced per absorption as a function of energy for fissile isotopes.

A-10 . . - - -...... , .

G, reactor power in megawatts, P, fraction of time Mechanical and Thermal Systems at rated power, f, and a as Design

(A-4) This section provides brief descriptions of core and blanket arrangements, fuel configurations, hence, vessel internals, heat transfer systems, shielding, 2.7 M instrumentation, and auxiliary systems. Discus- RDT = 0. sions relate to several design operations presently GPf (1 a) (A-5) + being used or under consideration for us(:. The doubling time is then measured in years and is proportional to the fissile specific inventory, Mo/P, Core and Blanket Arrangements. Two choices and inversely proportional to the breeding gain, G. exist regarding the arrangement of fertile and fissile Although Equation (A-5) is a simple model and has fuels in the reactor to optimize breeding potential. no provisions for time spent by the fuel in the fuel In the external breeding concept, all fertile material cycle outside the reactor, fuel cycle losses, fissions in is contained in the blanket surrounding the core; fertile material, and variations during the burnup hence, all breeding takes place external tci the core. cycle, it describes the sensitivity of the doubling time In the internal, or in-core breeding concept, some to the breeding gain and fissile specific inventory. fertile fuel pins are interspaced with fissile fuel More exact definitions of the doubling time, includ- within the core fuel economics. Figure A-3. illus- ing system or fuel cycle inventory doubling time and trates the two choices. compound system doubling time for equilibrium fuel cycles, are described in References A-2, A-3. The external breeding configuration results in a very hard spectrum, a low fissile inventory, and a good breeding ratio. Lack of in-core breeding Major Design Objectives results in a rapid reactivity loss with burnup, which requires frequent fuel changes, low burnup, a small doppler effect, few fast fissions in fertile material, a The principal goals of breeder reactor design are high fissile enrichment, and a thick bl;mket.A-4 safe operation, high breeding ratio, low doubling Thus, design of prototype and demonstration time, and low cost. Safe operations, in terms of LMFBRs allows for considerable internal or in- design objectives, prescribe a conscious awareness core breeding. of safety at every stage of design, and the attendant provision of reliable components and adequate safety margins for all normal and off-normal oper- Internal breeding core designs can be either ations. A high breeding ratio is needed to reduce homogeneous or heterogeneous. Figure A-4 shows fuel doubling time. Fuel doubling times must be top and side views of typical homogeneou!; and het- shorter than the doubling time of increasing electri- erogeneous core designs . Figure A-4a exemplifies a ' cal demand. Another important objective of any homogeneous core design because all of' the fuel energy generating system is low cost. Overall fuel assemblies containing pure fertile fuel are located cycle considerations are of paramount importance. in the radial and axial blanket regions. Therefore, a High burnup optimizes the use of high-fissile fuel relatively uniform or homogeneous mixture of fis- and minimizes down time for refueling. An addi- sile and fertile fuel is spread throughout the core. tional cost objective is the coordination of the The central region indicates the core itself, which breeder fuel cycle with that of the light water reac- contains the initial load of fissile/fertile fuel. The tor (LWR) cycle. Minimizing capital cost, which is outer region represents the radial blanket and directly related to hardware requirements, puts a shielding arrangement. The core typically contains premium on systems that require the least expensive two enrichment zones. The outer zone incorporates capital equipment commensurate with meeting the a higher fissile fuel content than the inner zone in criteria for a safe and reliable operation. The spe- order to effect a flatter radial power distribution. cific reactor design objectives required to accom- Control rods are normally boron-10, in the form plish these goals are listed in Table A-I. of B4C.

A-1 1 Table A-1. Goals and design objectives for the breeder reactor

Goals Accompanying Design Objectives

Safe operation Reliable components, adequate containment margin

High breeding ratio High neutron energy

Low doubling time High breeding gain, low fissile specific inventory

Low cost High burnup, compatibility with LWR cycle, minimal capital cost

a) External breeding concept b) Internal or in-core breeding concept

Figure A-3. External and internal breeding configuration.

A-12 control / Internal Radial shield1 blankets I

a) Homogeneous core b) Heterogeneous core

Figure A-4. Typical homogeneous and heterogeneous LMFBR core/blanket arrangements incorporating internal breeding.

In the heterogeneous core design, the blanket Fuel Lattice. There are great incentives to mini- assemblies, which contain pure fertile material, are mize the size of the LMFBR core. By reducing the distributed throughout the core region. The design coolant volume and structural material as niuch as yields a higher breeding ratio and reduced sodium possible, and thereby increasing the fuel volume void coefficients but requires higher fissile fuel fraction, neutron leakage is decreased, the fissile in~entories.~-~The heterogeneous core configura- fraction can be decreased, and reactivity can be tion (illustrated in Figure A-4b) contains blanket increased for a given amount of fissile material. assemblies at the center of the core and in concentric rings. Table A-2 compares 1000-MWe homoge- A triangular lattice arrangement intrinsically neous and heterogeneous designs .A-5 allows a higher fuel volume fraction than does a

A-13 Table A-2. Comparison of homogeneous and heterogeneous cores [IOOO-MW(e)designla

Homogeneous Heterogeneous

Number of fuel assembliesb 276 252

Number of in-core blanket assemblies - 97

Number of ex-core blanket assembliesC 168 144

Number of control assemblies 19 18

Number of enrichment zones 3 1

Fissile mass (kg) 3682 4524

Pu fraction by zone (wt%) 13.5/14.9/20.3 20.2

a. From Reference A-4.

b. 271 pins/fuel assembly, 1.22-m core height, 7.9-mm pin diameter, 0.36-m axial blankets.

c. 127 pindblanket assembly, 12-mm pin diameter.

square lattice. The higher fuel volume fraction min- normally hung from the reactor vessel, or reactor imizes fissile loading mainly by reducing neutron tank, as shown in Figure A-6. An LMFBR does not leakage from the reactor. The triangular lattice or require pressurization to keep the coolant in a liq- hexagonal structure has been selected for breeder uid state; outlet pressures are near atmospheric. reactor design. An LWR typically employs a square Consequently, the reactor vessel need only be thick lattice because relatively larger spaces between fuel enough to satisfy standard weight-bearing struc- rods are needed to optimize the water-to-fuel ratio. tural and safety requirements. A typical thickness The square lattice provides the space with relatively may be on the order of 30 mm, as compared to easy mechanical assembly. 300 mm for LWR systems.

Fuel Assembly. For the standard triangular pitch Assemblies are slotted into positioning holes in lattice, the pins are separated by a spiral wire wrap the core support structure, and radial core and assembled as a cluster of 217 pins within an restraints are normally provided at two axial loca- assembly Alternatively, grid spacers can tions. These locations are generally just above the be used to separate the fuel pins. Fuel pellets make active core near the top of the assemblies. Control up the active core region, and blanket pellets pro- rods enter from the top of the core, with the drive vide the axial boundaries. Fission gas plenums can mechanisms located atop the vessel closure head. be above or below the upper blanket region. The vessel normally hangs from a support ledge, Figure A-5 shows the geometric arrangement of the and the head is bolted to this structure. However, fuel, structure, and coolant. A table in the figure several variations in vessel support design have references the relationship between concentric been used. (hexagonal) rows of pins and the number of pins in an assembly. Sodium enters the lower inlet plenum, traverses up through the core and blanket regions, and col- Vessel Internals. The core of an LMFBR is lects in a large reservoir above the core before exit- located on top of a core support structure, which is ing to the intermediate heat exchanger (IHX). An n

A-14

C V (See enlargement below)

Row 2

Fuel assembly cross section

blanket ' Bottom - end cap Driver fuel pin Pin Pins Total Subassembly -row per row pins L 1 1 1 2 6 7 3 12 19 4 18 27 5 24 61 Fuel pin -7 6 30 91 diameter ' 7 36 127 8 42 169 ' ROY 2 y.,ROW 1 9 48 21 7 10 a 54 271 1/12 assembly cutaway 11 60 331 12 66 397

Figure A-5. Typical LMFBR fuel assembly system. inert gas, usually argon, separates the sodium pool pool system, the second the loop system. In the from the closure head. loop system, there is a sequential series of components between the reactor and the turbine:, each operating independently of the other loops in the System Heat Transfers system. Each loop consists of a single primary and secondary pump and one or more IHXs and sec- Pool and Loop System. Neutron activation of the ondary sodium system; there is no requirement that sodium coolant in the primary heat transport sys- the number of primary pumps and loops be the tem of an LMFBR requires that a secondary or same. Figure A-7 illustrates the heat transport sys- intermediate heat transport system also be tems for both the loop and pool designs. Table A-3 employed. lists the systems selected for different plant designs. LMFBRs using the pool design include EBR-11, Two different arrangements have been devised to PHENIX, PFR, BN-600, Super PHENIX, CDFR, accomplish this; the IHX and the primary pump and BN-1600. Those using the loop design include can be located inside the reactor tank, or they can DFR, Fermi, Rapsodie, BOR-60, KNK-2, IFBTR, be in adjacent hot cells with pipes connecting them JOYO, FFTF, PEC, BN-350, SNR-300, MONJU, to the reactor vessel. The first option is called the CRBRP, and SNR-2.

A-15 n Control rod drives

1 a n

Sodium inlet nozzle (3)

Figure A-6. Typical loop-type LMFBR vessel internals.

A-16 -

--111--- C 1 Tu rbi ne

system I Primary I I Pump 3 -r- 1 ------I

Steam f, generator

Core

Figure A-7. LMFBR heat transport system.

There are advantages to both the loop and pool transport and steam system. Whether thi,s is a net design. Loop designs include the following: advantage relative to pool design is still unclear.A-6

Maintenance is simpler because compo- Advantages of the pool design are as follows: nents can be isolated in cells. The feature also allows greater flexibility in making There is lower probability of a primary sys- system modifications and providing main- tem being broken tenance during reactor operation. Leakage of primary system components Less neutron shielding is required to pre- and piping does not result in leakage from vent neutron activation of the secondary the primary system outside the: reactor shielding. tank

The structural design of the vessel head is Sodium mass in the primary system is about simpler than the deck roof of a pool reactor. three times that in a loop design, 'thus providing three times the heat capacity The larger difference in vertical elevation of the IHX relative to the core in loop Temperature rise during off-normal tran- design enhances natural circulation over sients is lower and time to reach boiling is that of pool design. The difference allows longer if a heat sink is isolated more reliable prediction of the natural circulation in the primary coolant flow path. The pool's large thermal inertia tends to dampen transient thermal effects in other Because of the smaller mass of sodium in the parts of the system primary system in the loop design, there is a quicker response of the secondary sodium and steam sys- The cover gas system can be simpler, tems to changes in the reactor and the primary because the only free surface need8:d is that sodium system. This influences the control and in the tank, an exception being the super load following characteristics of the overall heat PHENIX, where an actively controlled

A-17 Table A-3. Main design parameters for large fast breeder reactorsa

Super BN-350 PHENIX PFR SNR-300 MONJU CRBRP BN-MX) PHENIX CDFR SNR-2 BN-I600 (USSR) (France) (U.K.) (Germany) (Japan) (US.) (USSR) (France) (U.K.) ~-(Germany) IUSSR)

Electrical rating, MW 350 (equiv) 250 250 327 280 375 600 I200 1320 I300 1600

Thermal power, MW 1000 568 600 770 714 915 1470 3000 3230 3420 4200

System LOOP Pool Pool LOOP LOO0 LOOP Pool Pool Pool LOOP Pool

Number of loops 6 3 3 3 3 3 3 4 6 4 4

Primary pump location Cold leg Cold pool Cold pool Hot leg Cold leg Hot leg Cold pool Cold pool Cold pool Hot leg Cold pool

Number of IHXs per loop 2 2 2 3 1 I 2 2 2 2 1

IHX temperatures

Reactor outlet, "C 500 560 550 546 529 535 550 545 540 540 550 Reactor inlet, "C 300 400 400 377 397 388 377 395 370 390 350 Secondary outlet, "C 450 527 540 528 505 502 520 525 510 510 505 Secondary inlet, "C - - - 335 325 344 - 345 335 340 310

Steam generator Integral; Modular; Separate; Separate; Separate; Separate; Separate; Integral; Integral; Integral; bayonet tube "S-shaped U-tube 2-straight helical hockey straight helical helical helical 9L m evaporator, I-helical coil stick tube coil coil coil or U-tube straight superheater tube

Steam cycle Recirculating Once-through Recirculating Once-through Once-through Recirculating Once-through Once-through Once-through Once-through - (Benson) (Sulzer) (Benson) (Benson) (Benson)

Number of units per loop

Integral steam generators ~ - - - - Separate evaporators 2 I2 I 3 I L Separate superheaters I 12 I 3 I 1 Steam drums 0 I I 0 0 1 Moisture separators - I - I 1 - Reheaters 0 12 I 0 0 0

Turbinc

Inlet pressure, MPa 4.9 16.3 12.8 16.0 12.5 10.0 14.2 18.4 16.0 16.5 14.0 Inlet temperature, "C 435 510 513 495 483 483 505 490 490 495 500 Number/rating, MW(e) I I 11250 I 1 1/434 - 2/600 2/660 1/1300 2/R00 Type K-100-45 Condensing Tandem comoound Condensing Tandem comoound Tandem comvound K-200-130 Condensing Tandem compound Single shaft -

a. From Reference A-8. cover gas is used to reduce reactor vessel Figure A-9 illustrates four steam cycles, modifi- wall thermal stresses. cations of the basic Rankine cycle, being considered for LMFBRs. Steam cycle. LMFBR core inlet and outlet temperatures are generally about 400 and 550°C, The Benson cycle is a once-through superheated respectively, and correspond, except for small heat cycle, employing an integral steam generator. The losses, to the temperatures across the primary side Sulzer cycle is another once-through superheated of the IHX.A-7 cycle but it uses separate evaporators and superheaters with a moisture separator between them. Primary and secondary flows in the IHX are gen- The exit quality from the evaporator in the: Sulzer erally countercurrent, with log mean temperature cycle is high. For the SNR-300, it is purported to be differences between the primary and secondary 95%.A-93-10 CRBRP design states that exit quality sodium on the order of 30 to 40°C.A-7 With the from the evaporators in a recirculation superheater exception of the PFR, secondary flow is on the tube cycle is 50%. side in order to facilitate cleanup of sodium-water reaction products from the steam generator, should Plant Control. The factors involved in the control a leak occur Pressure on the secondary of an LMFBR heat transport system can be (derived side is higher than on the primary side to avoid from the following brief outline of the plant control leakage of radioactive sodium to the secondary, in system developed for the CRBRP. Figure A-10 is a the event there are tube leaks. schematic of the two-level CRBRP plant control system. The systems provides automatic and man- An additional reason for placing the secondary ual control of the reactor, heat transfer system, tur- on the tube side is that components of the heat bine, and auxiliary systems for both normal and transport system in both the primary and second- off-normal operation. The first level of control is ary system are arranged such that the thermal ten- the power load demand, which is established by ter of each component is above the thermal center either an automatic signal from the power grid load of the previous component in the flow cycle. dispatching system or a plant-operator-determined setpoint. The supervisory controller receives the Steam generators can be categorized as integral load demand signal and data on steam tempera- or separate. In an integral system, evaporation and ture, steam pressure, and generator output. The superheating occur without separation of the water supervisory controller sends signals to each of the and the steam between the two processes. Many second-level controllers, which brings about any integral steam generators employ both evaporation adjustments required in the reactor power level, and superheating within an integral unit (i.e., system flows, and steam supply to match the power within the same shell). An example is the steam gen- demand. erators to be used in the Super PHENIX. Other Each of the second-level controllers can be oper- integral steam generator systems employ separate ated manually. Manual control is normally used components, as in the BN-350. Figure A-8 illus- when the plant is shut down, and during startup to trates examples of integral steam generators. bring the plant to a power level at which the automatic control system can be initiated. The auto- In a separate steam generator, evaporation and matic control system normally takes over at about superheating occur in different units, with steam 40% of maximum plant power. separation between the two processes. Steam separation usually takes place in a steam drum or mois- The reactor controller receives a demand1 input ture separator, which are separate components regarding reactor power level from the supervisory between the evaporator and superheater. Steam controller and input of sodium temperature arid neu- separation can be incorporated as an integral part tron flux level in the reactor. The controller coinpares of the evaporator. Steam-water is always on the the flux level with the power demand signal to deter- tube side, sodium on the shell side. The steam- mine the change required to meet the demand. A trim water pressure is greater than the sodium pressure. signal from the steam temperature sensors is sent to In the event of a tube leak, steam or water will flow the turbine throttle to maintain throttle conditions at into the sodium. This will prevent contamination of a given temperature value. At the same time, in forma- the turbine with sodium oxide. tion of core outlet temperature is used to modlify the

A-19 Figure A-8. Plain view of Super Phenix pool, showing the four loops, four primary pumps, and eight intermediate heat exchangers. controller output so that core outlet temperature is from the pump controllers to the pump drive con- maintained within predetermined limits. The result- trols for any changes in pump speed. ing signal is a demand to the control rod drive mechanism controller to adjust the rod position to a point The steam drum and feedwater controller does where the difference between power output and not receive input from the supervisory controller demand is minimized. because the steam drum level remains constant for all power levels. The controller receives inputs of The primary and intermediate sodium flow con- flow in the main steam line, steam drum level, and trollers receive inputs of flow demand from the flow in the feedwater line. Comparison of these supervisory controller and sodium flow in the cold inputs result in a demand signal to the feedwater leg of each loop. The controller automatically com- control valves, which adjust flow to maintain steam Dares the demand with the flow signal to determine drum level. if any changes are required. A trim signal is provided to the intermediate pump flow controller The turbine generator controller receives inputs from the turbine throttle to maintain throttle pres- of power load demand from the supervisory con- sure at the preset value. The resulting signal is a troller, steam flow temperature and pressure, tur- demand to the pump controllers, which also receive bine speed, and generator output. Flow, input from tachometers on each of the pumps. temperature, and pressure signals are combined to Comparison of the inputs result in a demand signal give mass steam flow information. The generator

A-20

- Integral superheater evaporator

Feedwater Feedwater heaters heat e rs

Once-through superheated Once-t hrough superheated (Benson cycle) (Sulzer cycle)

Superheater

Steam

Evaporator

Feedwater Feedw at e I' heaters Iheaters Recirculating superheated cycle Saturated steam cycle

Figure A-9. LMFBR steam cycles. output signal is compared to the power load ture is attached to an inner guard vessel (not present demand to establish any requirements for power in CRBRP), which in turn is supported at the bot- change. Comparison of these demands results in a tom by the reactor vessel. In the CRBRF' design, signal from the turbine generator controller to the the core support structure is attached to a tore sup- turbine throttle valve, which adjusts steam flow to port ledge, which is attached to the side of the ves- meet the power load demand. The turbine speed sel. In both SNR-300 and CRBRP, a core barrel or signal is used to trim for small variations in turbine jacket is joined to the core support struci.ure and speed. separates the sodium flowing through the core, blanket, and radial shielding from the surrounding Components sodium pool. An inlet flow structure guides the sodium from the inlet plenum to the fuel assem- Reactor Vesseland Reactor Tank. In a loop design, blies; an upper internal structure guides 1 he flow the reactor vessel is a vertical, cylindrical shell with from the assemblies into the outlet plenum. Con- a dome-shaped bottom. The vessel is hung at the tainers for interim fuel storage during refueling are top from a support ring. The fuel assemblies rest on located outside the core barrel. a core support structure. Figure A-1 1 illustrates an example of a reactor vessel for a loop system, the A guard vessel is located outside the reactor ves- SNR-300. In the SNR-300, the core support struc- sel to protect against any potential loss of sodium

A-2 1 1st level control

2nd level :LSteam press control

Primary flow

Intermediate- 0TIC Thermocouple -1 flow 0TIG Tachometer

Figure A-10. The Clinch River Breeder Reactor project design for plant control. Control roc1 drives

Rotating plugs

Fuel charging~ channel (outside vessel) Sodium inlet Submersed cover plate Instrumentation plate

GuGd v

Reactor~ ~ ve5;sel Inner guard1 vess el Core barre I Core diagritl

Guide vanes

Lower collector vessel

Figure A-1 1 . Reactor vessel for the loop system of SNR-300.

from the vessel. The reactor and guard vessel are set The Super PHENIX reactor tank design is shown in a reactor cavity. Both inlet and outlet sodium in Figure A-12. The tank is stainless steel, has a pipes enter above the guard vessel so that any pipe height and diameter of 19.5 m and 21 m, respec- rupture within the reactor cavity will not result in tively, and a wall thickness of approxiinately sodium loss to the core. In the SNR-300, the inlet 50 mm. It is hung from the deck that covm the pipe penetrates the vessel above the outlet pipe, and tank. runs down between the inner guard vessel and the reactor vessel to the inlet plenum below the core. In Further details of the reactor tank and ini:ernals the FFTF and the CRBRP design, the inlet pipes are shown in Figure A-13. Asafety tank surrounds run down between the reactor vessel and the guard the reactor tank to contain sodium in the event of a vessel and enter the vessel near the bottom. leak. A baffle tank inside the reactor tank prevents the tank from exceeding the sodium reactor inlet Several centimeters of argon gas cover the temperature, accomplished by providing by-pass sodium, separating the sodium pool from the reac- sodium flow from the inlet plenum to the aiinular tor head. The reactor head provides access for both space between the baffle and the reactor tanks. control rod and refueling ports and is usually fabri- Another internal tank is the insulated internal cated with several rotating plugs for refueling; tank, which forms both a physical and thermal barrier between the hot and cold sodium. The penetra- The reactor tank for a pool system is described tions for the primary pumps and IHXs are by using the design adapted for Super PHENIX, supported from the deck. Because of the large tem- the first commercial-sized pool reactor to be built. perature variations, from the room temperature

A-23 Control rod drives n

Figure A-12. Reactor tank for the pool system of Super Phenix. above the deck to the sodium pool temperature supports the tank. The deck must also support pri- below, the IHX penetrations of the insulated inter- mary pumps, IHXs, control rod drive, and fuel nal tank must accommodate 50 to 70 mm of axial handling equipment, and must, in addition, be expansion. A bell-jar type seal is used to accom- designed to remain leak-tight following potential plish this, as shown in Figure A-14. The pump fits deformation from a hypothetical core disruptive loosely into a well that extends from the cold accident. As in the case of the loop system, the deck sodium to the deck so that no seal is required. is separated from a sodium pool by a cover gas. In Another problem for the pumps, however, involves addition, the deck must be insulated from the high- the pipe connection to the inlet plenum. A flexible temperature sodium. Even with insulation, some joint must be designed to accommodate the cooling of the deck is required. Insulation for the approximately 50-mm radial and 100-mm axial Super PHENIX is stainless steel gauze sandwiched thermal expansion created by differences between between stainless steel foil, Similar insulation is the deck and pump operating temperatures. also required to insulate the concrete walls surrounding the reactor tank. Construction of the deck is' illustrated in Sodium Pumps. Sodium pumps in general use at Figure A-15. It is composed of a steel web filled LMFBRs are normally mechanical, vertical shaft, with concrete and is supported by the concrete vault single-stage, double suction impeller, free surface, that surrounds the reactor tank. The deck in turn centrifugal pumps.

A-24 l In a pool system, the primary pump i,; always located in the cold sodium. In a loop system, however, location of the primary pump can be either on the hot or the cold leg. The pump is located in the cold leg of all secondary systems because the advantages of a hot leg location existing for the primary loop are not present and also because it is important to pressurize the secondary sodium in the IHX in order to force flow from leaking tubes in the direction from the nonradioactive secondary to the radioactive primary.

In the loop design, there are some obvious advantages to seals and bearings, for exarnple, of placing the primary pump in the colder sodium environment of the cold leg. This was Idone in MONJU, BN-350, and most of the early experimental loop reactors. The hot leg has been selected for several loop designs, SNR-300, SNR-2, FFTF, and CRBRP. The main reason for choosing the hot leg involves suction requirements for primary system pumps. These requirements are stringent to accommodate the full range of anticipated transients. The basic argument can be explained briefly, however, by examining conditions during normal steady state operation.

Figure A-13. Details of the Super Phenix reactor tank The net positive suction head (NPSH) of a pump and internals. is the difference between the absolute pressure of the pump suction and the vapor pressure of the Choices affecting pump design involve differ- fluid pumped. Any pump has, inhereni. in its ences between primary pumps for the loop and design, a minimum required NPSH at any given pool systems, minor differences between primary flow rate, to prevent cavitation. The available and secondary pumps, and pump location. Impor- NPSH must, therefore, always be greater than this tant design choices include seal, bearings, impeller, minimum value. The available NPSH can be and by-pass flow arrangements. obtained from the following relationship:

Electromagnetic pumps have also been used for NPSH = H + H, - HI - Hv ('4-6) LMFBRs, because sodium is an excellent electrical P conductor. An electromagnetic pump is being used where for the secondary sodium in EBR-11, and they have been used in SEFOR, in the primary circuit of the pressure at the liquid surface of the Hp = Dounreay fast reactor, and in the SLSF test train at source from which the pump takes ETR in Idaho. They are not used in the main loops suction, in this case the cover gas pres- in large LMFBR power plants, but they are being sure used in some back-up decay heat removal systems, for example in SNR-300 and Super PHENIX. H, = hydrostatic head of the liquid source above the impeller Only mechanical pumps are being used in large LMFBRs. Testing is underway in the United States H1 = pressure drop losses in the piping and on the use of a relatively small inducer-type equipment upstream of the pump mechanical pumping element in series with a large centrifugal pump to assist in meeting the strict suc- Hv = vapor pressure of the fluid at the suction requirements of an LMFBR centrifugal pump. tion.

A-25 -1-

operation temperature I

Figure A-14. Super Phenix’s intermediate heat exchanger penetration into an insulated internal tank.

IHX or pump penetration ’\\

\ Steel skin

girders

Radial beams

Figure A-15. Construction of Super Phenix’s roof/shield deck n

A-26 The pressure above the sodium surface in sion, used in FFTF, is to design a bend in pa.rt of the LMFBR primary pumps is equalized with the reac- tube bundle. A sine wave bend is used in E’FR and tor cover gas pressure, H P’ and need be only CDFR. slightly above atmospheric, just enough to prevent leakage of air or inert gas. The hydrostatic head, Each of the eight IHXs at Super PHENIX includes H,, controls the length of the primary shaft. The 5380 tubes, 14-mm outside diameter, 12-min inside sodium vapor pressure, H,is small, on the order diameter, and 6.5-m length. A remotely operated of only 1 kPa for the hot leg sodium. machine has been developed to seal off tubes that leak. Both 316 and 304 stainless steel are generally If the pump is on the hot leg, the pressure drop used in the IHX components. Hi includes only the losses through a short distance of the piping, plus vessel exit and pump entrance Steam Generators. AS discussed above in the losses. With the pump on the cold leg, however, HI “Steam Cycle” section, steam generators can be must include the pressure drop through the IHX, integral or separate. They can use tubes #that are plus additional piping. The IHX pressure drop is straight, helical, U-tube, or hockey-stick shaped. generally on the order of 50-100 kPa. To obtain a Either single- or double-wall tubes can be used. NPSH in the cold leg comparable to that in the hot Table A-3 presents steam generator characteristics leg would require an increase in the length of the for the prototype and demonstration plant:;. pump shaft of H,, or to pressurize the argon cover gas in the reactor, Hpor a combination. Pressuriz- The choice of integral versus separate steam gen- ing the cover gas would increase the potential for erators involves the selection of steam cycle (dis- leakage of radioactive gas through the cover seals. cussed in the “Steam Cycle” section above). Another option is to develop a more advanced Figure A-17 portrays the general shape of the tem- pump that can operate with a lower NPSH. The perature distributions in an integral unit. After difficulty of achieving these options is the reason boiling is complete, the steam can be superheated for the hot leg location. The problem is further close to the inlet sodium temperature. aggravated as LMFBR plant size increases, because the required NPSH for a particular pump increases In separate steam generators, the subcooled within increasing rated capacity. heating and boiling take place in the evaporator, whereas the superheating occurs in the superheater. Intermediate Heat Exchangers. IHXs at all proto- Complete evaporation does not occur in the sepa- type and demonstration plants, with the exception rate evaporator. CRBRP evaporators, exemplifying of the BN-350, which uses U-tubes, are vertical the recirculating cycle, were designed for 50% exit counterflow, shell and tube heat exchangers with quality. Exit quality for the SNR-300 evapcrators, basically straight tubes. Figure A-16 illustrates an which operate on the Sulzer once-through cycle, is IHX for both a pool design (Super PHENIX) and a 95%. In both cases the steam is separated from the loop design (CRBRP). With the exception of PFR, liquid before entering the superheater. As can be the secondary sodium enters at the top and flows to observed from Table A-3, both integral and sepa- the bottom through a central downcomer; flow is rate systems have been widely used. then reversed and returns up through the tubes. Pri- mary sodium generally flows down on the shell side Several designs for steam generators have evolved, and exits at the bottom. Principal differences in all having particular advantages and disadvantages. IHXs between the loop and pool system appear at A fundamental consideration in each design is the the entrance and exit flow nozzles. method for accommodating thermal expansion. Fig- ure A-1 8 presents examples of the basic configura- The tube bundle must be mounted to allow for tions capable of providing such accommodations: differential thermal expansion between the tubes helical coil, U-tube, and hockey stick. The figure also and the shell. To accomplish this, the lower tube shows the straight tube configuration, which requires sheet is allowed to float, and, therefore, is sup- special provisions to accommodate thermal expan- ported by the tube bundles. The tubes are sup- sion, similar to those described for the IHX. Helical ported by the upper tube sheet. The CRBRP design coils are used in Super PHENIX, CDFR, SNR-2, and shows a flexible bellows at the top of the down- MOJNU. U-tubes are used in PFR and in the EIN-350 comer to allow differential expansion between the superheaters. Straight tubes are employed in BN-600, downcomer and the tube bundle. An alternate as well as in two loops in SNR-300. A third loop of

,\ method for accommodating the differential expan- SNR-300 incorporates a helical coil steam generator.

A-27 n

C 0 L al n

*v) :: a 0 -0 a e? m ud

0 0 a

WI.. a rns W 5 0 Y I.. al C P U ca L E Q, .a, , C .-X r .- .-C 5!?% C c-ox

A-28 of austenitic stainless steel for the original PFR superheaters and reheaters.

An important condition in steam generator design is the transition between nucleate boiling and film boiling, or the departure from nucleate boiling. At this point, the temperature of the tube wall rises sharply and an instability in tube wall temperature occurs in the transition zone. This temperature behavior is illustrated schematically in Figure A-19. The transition occurs between X1 and X2. In this region, 0 x2 x3 the tube wall is intermittently in contact with water or steam, and the wall temperature fluctuate rapidly. Figure A-17. Temperature distributions in an integral Such fluctuation, if too large a magnitude, can cause superheater. thermal fatigue of the tube, or structural changes that enhance water-side corrosion. Either helical coils or straight tubes will be used in SNR-2. In the United States, the hockey stick was Shielding. Requirements for shield design are selected and tested for the CRBRP. Design and devel- greater for fast reactors than for thermal reactors. opment of both helical coil and straight tube designs The high energy neutron flux in an LMFI3R is con- are also underway. Another type of tube is used in the siderably higher than in an LWR, as is the high- BN-350 evaporator called bayonet tubes. These tubes energy neutron leakage from the core. Although have a central tube surrounded by an annulus. Water the neutron production rate for an LMFBR and flows down the central tube and a water-steam mix- LWR have the same power level, or are cornparable, ture flows back up the annulus. the power density (kW per liter) is higher for the LMFBR. Also, the neutrons in an LWR are slowed Tube integrity is far more important for the down to thermal energies close to the fission LMFBR steam generator than for the LWR because source, so that the high-energy leakage source for of the potential chemical reaction between sodium radiation of surrounding structures is relatively low. and water. Early consideration was given to the use of double-walled tubes in which leaks could be LMFBR shielding can be discussed according to detected from gas conditions between the two areas that require extensive shielding analysis and walls. Double-walled tubes are used, for example, design. For both pool and loop designs, these areas in EBR-I1 and Dounreay fast reactor steam genera- include in-vessel radial shielding, the closure head tors. However, when experiments demonstrated assembly and its penetrations, and neutron flux that sodium-water reactions could be adequately monitors. For pool designs, special attention is contained, the emphasis shifted to the simpler required for the intermediate heat exchangers, single-wall tube design, and all prototype and dem- where secondary sodium can be activated by neu- onstration plants now use a single-wall tube. Some trons. For loop designs, other key areas include the interest in double-wall tubes is being revived, but reactor vessel support area and the primary heat the argument that they might lead to greater steam transport system pipeways. Other areas that require generator reliability was more especially important shielding include auxiliary piping penetrations, for the early generator plant. Whereas experience I heating and venting system penetrations, shielding with LMFBR steam generators has been reasona- for cover gas and coolant purification systems, fuel bly encouraging, BN-350, PFR, and the Fermi handling equipment shielding, and biological reactor have all experienced difficulties with leaks. shielding for areas that require personal access during normal pool operations. Most steam generators, both tube and shell, are made of ferrite steel containing 2.25% chromium Coolant and Cover Gas Radioactivity. Thi: use of and 1% molybdenum, selected to minimize chlo- sodium as coolant in the LMFBR inh-oduces ride stress corrosion. In some cases the material is shielding problems different from those in the stabilized with 1070 niobium to reduce carbon loss LWR, as a result of neutron activation of the to the sodium. Exceptions to this are the selection sodium. Sodium in nature is composed eritirely of of Incalloy-800@ for Super PHENIX, and the use sodium-23. The (7 , y) reaction in sodium ~xoduces

A-29 Water inlet

Steamlwater outlet rs/ f?

Closure head

Sodium outlet Sod iumlwater

Sodium inlet Sodium inlet from reheater from superheater

-%&-Sodium mixing chamber Sodium-- drain-- ---4

Figure A-18. Steam generator designs.

A-30 Argon

Sodium inlet

- Rupture disk (to separator tank)

,Bleed vent

Water inlet

Sodium - 'Tubes

-,A:..-

/ Shell

Tube spacer Sodium outlets YIw

Hockey stick (C R BR P) Inspection openiiig and sodium drain1

Water I inlet Figure A-18. (continued)

A-3 1 presents calculated activity caused by fission gas in the reactor cover gas for these design base conditions. Since failed fuel will likely never approach Q 1070, actual activities will likely be far below these levels.

In- Vessel Shielding. In-vessel radial shielding is required to prevent excessive radiation damage to structural materials that must remain in a reactor for the lifetime of the plant, and to protect the ves- 1 1 I 1 sel itself. Examples of in-vessel structures are the x1 x2 core barrel and the core restraint systems. Axial shielding below the core is needed to protect the Heated tube length core support structure. The shielding design must in the boiling section ensure that permanent components have an end-of- life ductility consistent with the threshold criterion Figure A-19. Instability in rise in wall temperature at the for brittle fracture. In both FFTF and CRBRP transition from nucleate to film boiling. design, the threshold of ductility is chosen to be 10% total elongation, the level that ensures a duc- radioactive sodium-24, which has a 15-hour half- tile mode of deformation up to failure, and permits life, and emits both a 1.4- and 2.8-Mev gamma with conventional structural analysis methods and crite- decay. A (77 ,277) threshold reaction also occurs, ria to be used in design. producing sodium-22, which has a 2.6-year half- life, and emits a 1.3-MeV gamma. During opera- The radial blanket serves as the first shield tion, sodium-24 is the dominant activation between the core and the radial structure. FFTF has product, and shielding against gammas from this no blanket but uses removable radial reflectors to source in the primary sodium is one of the impor- protect the fixed radial shield and the radial sup- tant shielding problems in LMFBR design. port structure and core barrel. Beyond the blanket Sodium-22 activity becomes the dominant activity are located removable radial shielding (RRS) in the sodium approximately 10 days after shut- assemblies. RRS assemblies contain rods of stain- down. For maintenance of primary pumps and less steel or of nickel-based alloys compatible with IHXs, however, radioactivity from corrosion prod- sodium, such as Inconel. They have a high inelastic ucts becomes the most important radioactive scattering cross section and are particularly effec- source. The calculated sodium-24 specific activity tive in degrading the energy of fast neutrons. Iron is in the CRBRP primary sodium is 30 Ci/kg (based also effective, but less effective than nickel. Stain- on a primary system sodium inventory of less steel is less expensive, however, than Inconel. 6.4 x lo5 kg). For FFTF, the calculated sodium-24 The choice of material for CRBRP was narrowed to activity is 11 Ci/kg. For an early (1968) General 316 stainless steel and Inconel-600. The shielding Electric pool design, the calculated sodium-24 rods in the CRBRP design extends from the bottom activity was 18 Ci/kg, based on a rimary system of the lower axial blanket to the approximate top of sodium inventory of 1.3 x 106 kg. 1-11 Calculated the upper axial blanket. The shielding is in the form CRBRP sodium-22 specific activity, after 30 years of rod bundles in order to allow cooling of the of operation is 3.5 mCi/kg. The corresponding shield by sodium. The main heat sources in the FFTF value for sodium-22 is 1 mCi/kg. shield are gammas from the core and blanket, and gammas generated in the shield itself both from Another radiation source in an LMFBR is the inelastic scattering and neutron capture reactions. reactor cover gas. Activation of impurities in the In FFTF, the radial reflectors are made up of hexag- sodium, and direct activation of argon 40 to onal blocks of Inconel bolted together. Holes argon 41, contribute to activity in the cover gas. through the blocks provide cooling. Neon-23 appears from an (q ,p) reaction with sodium-23, but its half-life is short (38 seconds). In CRBRP design, the RRS assemblies are sur- The main design requirement, however, is to permit rounded by a fixed radial shield. The fixed radial reactor Gperation with leakage occurring in a speci- shield is an annulus of 3 16 stainless steel, 0.146 m fied fraction of the fuel pins. For FFTF, this frac- thick, which can experience relatively high fluence tion of defective pins was set at 1%. Table A-4 because it is not a loadbearing component. The

A-32 Table A-4. Calculated cover gas activity for the FFTF design basis, 1% pin failure Ir) Activity Isotope (~i/m3)

Kr-83m 6.82~lo1

Kr-85m 1.34 x lo2

Kr-85 9.30 10-9

Kr-87 1.80 x lo2

Kr-88 2.64~lo2

Xe-133m 5.43 x 10-1

Xe-133m 1.47 x lo1

Xe-133 2.67 x lo2

Xe-135 1.26 x 103

FFTF fixed radial shield is formed from flat plates Reactor Enclosure System Shielding. Areas requir- to create a 12-sided shield between the core and the ing shielding of the reactor enclosure system core barrel. include component penetrations and interfaces in a closure head assembly, the reactor vessel support In the pool design, graphite is usually incorpo- area, and the ex-vessel flux monitors in the reactor rated into the radial shielding to moderate neutrons cavity. Figure A-20 illustrates these areas for and allow them to be absorbed before reaching the CRBRP, which has a typical loop design. For a IHX. B4C shielding is used near each IHX and pri- pool design, neutron flux levels at the tank support mary pump in order to reduce activation of the area and reactor cavity walls are lower than for the structural material, thereby allowing maintenance loop design, so shielding problems in thest: areas as well as reducing secondary sodium activation by are less severe. thermal neutrons in the IHXs. The CRBRP design solutions in these areas indi- An axial shield is located below the lower axial cate the types of shielding problems encountered. blanket to protect the core support structure and lower inlet modules. In CRBRP, this shield consists Figure A-21 shows the CRBRP closure head of a 0.51-m-long, 316 stainless steel shield block in assembly (CHA). The main penetrations invohe refu- each fuel, blanket, and control assembly. In FFTF, eling components, control drive mechanisms, and the a lower shield and inlet orifice assembly is 0.54 m upper internals lifting mechanism. Radiation source long. No special upper axial shielding is required terms that affect the CHA shielding design iiiclude for either FFTF or CRBRP because of the shielding neutron and gamma streaming of the step annuluses provided by the upper sodium pool. of the CHA penetrations or component interfaces, the radioactive cover gas below the CHA pmetra- Pathways for neutron streaming from in-vessel tions, and neutron and gamma penetrations through components in CRBRP include the clearance gaps the CHA bulk shielding. required in the design of the fixed radial shield, the cooling channels of the axial shielding set in each core The sodium dip seals around the refueling ]plugs, assembly, the fission gas plenum of each core assem- shown in Figure A-21, presents the major shield bly, and interfaces between invessel components. design problem for the closure head assembly.

A-33 vessel Closure head area and assembly ledge

Reactoi

Figure A-20. Key shielding areas in CRBRP reactor and reactor enclosure system.

These seals form the barrier for the cover gas in the tion of the support ledge to stop thermal neutrons, CHA rotating plug annuluses. CRBRP was and a carbon steel collar reduces the streaming gap designed to operate with 1Yo failed fuel. Radioac- at the vessel flange elevation. A concrete shield ring tive fission product gases and the cover gas from above the support ledge reduces radiation stream- this failed fuel requires about 0.3 m of steel shielding into the head access area. ing for personal access to the head access area. Dip- seal tradeoff studies resulted in the location of the The source range flux monitor (SRFM), which seals in the closure head, as shown in Figure A-21. monitors the core during shutdown and refueling, Radiation through these dip seals is the largest con- is located in the reactor cavity of the CRBRP. The tributor to the dose rate in the head access area. neutron flux from the core that reaches the flux monitors must be great enough to allow monitoring In both FFTF and CRBRP designs, a canned of changes in core subcriticality, whereas neutron B4C radiological shield is placed at the lower eleva- fluxes from extraneous sources and gamma dose

A-34 Upper internals Control rod drive structure support Intermediate [mech :-ism n-OkZ!.g column nozz!e \ rotating plug Liquid level monitor Ex-vessel tran achine nozzle

plate column

Figure A-21. Closure head assembly configuration of CRBRP.

rates must be sufficiently low. These objectives are tion of gamma rays from sodium-24 in the primary accomplished by shielding at the SRFM. The coolant pipes with deuterium in the concrete. In the shielding consists of a graphite moderator block, CRPRP design, more than 80% of the no&u t ron 0.51 m by 0.63 m, surrounded by lead and B4C background at the delayed neutron monitors could background shields. Neutron background, caused be attributed to photoneutrons from the concrete. by fuel-in-transfer or fuel storage in the fuel trans- Nonhydrogeneous materials were specified for neu- fer and storage assembly, is reduced by B4C shields tron background shielding around the moniiiors to in the reactor cavity. The gamma background at the minimize the neutron flux levels. SRFM is reduced to acceptable levels by surrounding the moderator block with lead to reduce gamma Instrumentation. All nuclear reactor systems levels from the vessel, guard vessel, and sodium, incorporate a high degree of instrumentat ion in and by using a high-purity aluminum alloy as the order to provide continuous monitoring necessary structural material for the SRFM to minimize its for plant control. Much of the instrumentation, neutron activation gamma background. such as radiation monitoring, is common to all reactor types. The presence of a liquid metal., how- Heat Transport System. The intermediate heat ever, poses a few instrumentation challenges unique exchanger in the loop design system must be to LMFBR systems. shielded from neutrons in order to prevent activation of the secondary sodium. In CRBRP, second- COR rnrameter Monitoring. Variables such a:, flux, ary sodium activation is held below 0.07 mCi/kg. temperature, flow, and pressure must be determined Ordinary concrete structural walls of the equip- in any nuclear reactor, but the sodium environment ment cells provide the bulk shielding. Considerable requires measurement techniques somewhat different design effort was required to reduce neutron from those of light water systems. streaming through the pipeways into the cells. Delayed neutron monitors for detecting fuel clad- It was noted in the section on “Reactor Elnclo- ding failure were also placed in the CRBRP heat sure System Shielding” above that flux monii.oring transport piping system; hence, the background for a typical LMFBR system consists of a number neutron flux levels at these monitors must be mini- of neutron detectors located in the reactor cavity mized. A problem encountered in shielding design external to the reactor vessel. In-core or in-vessel is the photoneutron production in the concrete cell detectors may be used for initial start-up operation, walls. Photoneutrons are generated by the interac- but the neutron source from spontaneous fission of

A-35 plutonium-240 is normally present and is strong magnetic flowmeter. It is possible to calibrate both enough to activate such remotcly located detectors. types of meters by activating the sodium with a Q An appreciable neutron source is also possible from pulse neutron device and using a time-of-flight americium-242, especially if recycled fuel is used. recording technique. The procedure was successfully employed in FFTF. Figure A-22 shows a flux monitoring set for the CRBRP design. The highly sensitive BF3 detectors Another type of flowmeter made possible by the are used for low-power operation, such as start-up unique properties of the liquid metal is an eddy cur- and refueling. Uranium-235 fission chambers are rent flowmeter. Figure A-25 is a diagram of such a employed for mid-range operations. And compen- device used in measuring the rate of coolant dis- sated ion chambers are used for power range mea- charge from an assembly into the upper sodium surements. Three sets, identical to that shown in pool. Figure A-22, provide an overlapping range of flux monitoring to continuously record neutron flux from shutdown to more than full power. Electrical Liquid pressure measurements are normally made signals from these detectors, which are propor- by routing a small column of the high-pressure liquid tional to reactor power, are used for both reactor onto one side of a sensing diaphragm. This causes a control and the plant protection system (PPS). complication when measuring sodium pressure Such signals also feed the data logging system and because sodium solidifies well above room tempera- provide annunciator trips from the control room ture. Trace heating could be provided to ensure liquid for out-of-limit conditions. sodium conditions, but this has become unreliable for many applications. An alternative method often Sodium temperatures must be measured rou- employed is to interface the sodium with NaK via a tinely throughout the primary and secondary cir- bellow system, as illustrated in Figure A-26. The NaK cuits to calculate thermal power and determine is a liquid at room temperature and can be used in a loop operating conditions. Two types of detectors close tolerance pressure transducer assembly. are in common use: resistant temperature detectors (RTDs) and thermocouples. The RTDs provide Fuel Failure Detection. The detection of a cladding a highly accurate and reliable measurement to breach in a fuel element can be normally accom- ensure that the plant is operating within design lim- plished by monitoring increased cover gas activity its. The sensor for such a device typically consists or by detecting the presence of delayed neutrons in of a double element of platinum contained within a the sodium leaving the reactor. Locating the fuel sheet that is spring-loaded against the bottom of a assembly containing the leaking fuel element is thermowell, as illustrated in Figure A-23. Insertion more difficult. Gas tagging offers one means for of this sensor raises the possibility of a sodium such identification. leakage from the penetration in the event of thermowell failure. Whereas well failures in such devices appear to be rare, a backup cable penetra- A cover gas monitoring system normally exists to tion seal (as shown by the seal connection pad in indicate the presence of fission products that Figure A-23) can prevent such leakage. escape the fuel pin. The most abundant fission products that escape are isotopes of the noble Coolant flow measurement must be made to gases, xenon and krypton. Many of the fission complete the thermal power calculations and loop products emit gamma rays with relatively low operating characteristics. Both the standard venturi energy, around 100 KeV, and detection of such flowmeter and a magnetic flowmeter are often used activity in a prevailing background of high energy on liquid metal systems. The magnetic flowmeter is neon-23 (440 KeV) and argon-41 (1300 KeV) gam- unique to liquid metal systems owing to the electric mas presents difficulties. The neon-23 activity is properties of the liquid metal coolant (see considerably less in a pool type reactor, relative to a Figure A-24):The Venturi meter is highly accurate loop-type system, because of the larger holdup but suffers from a response time that is often too time, which allows the 38s half-life neon-23 to slow for control system and PPS use. The magnetic decay, and less turbulence in the pool level. How- flowmeter, on the other hand, tends to be less accu- ever, numerous xenon isotopes emit relatively high rate but exhibits rapid response. When the two are energy gammas, and experience at EBR I1 and used in series, the Venturi meter can be used to FFTF indicates detection of cladding failure for an n achieve in-place calibration of the rapid-response LMFBR system is readily attainable.A-12,-13

A-36

1. Plant protection Other systems

4- Flux monitoring system - -8 s ys t e mh---*

BF3 detector Preamplifier I Signal Bu f f e rs - condition - - A Annitnciation (typical of 3 ex-vessel c han ne I . LcrData hand I ing U-235 fission Preamplifier chamber I , -, F1Display

I Annunciation

c I Primary ' DC linear I - Junction box - shutdown .ion chamber I power ' logic

k1Data handling

Figure A-22. Flux monitoring set for CRBRP design.

A-37 n

Sealed connection head ,Instrument tree dry well

Union ddy current flowmeter probe

:.".:, .:...: .' .....':I )--( p-

Flow guide tube IlQ9Pf u Flowing sodium

Figure A-24. Magnetic flowmeter and flow tube for large sodium coolant piping.

Figure A-23. A typical thermowell assembly incorporating a resistance temperature detector (RTD) for measuring sodium temperatures.

n Figure A-25. Simplified diagram of eddy current flowmeter.

A-38 NaK filled capillary tube

Sodium to NaK interface bellows seal assembly h NaK - Sensing diaphragm

r ' .. .: :..: . . . ., .. . i :. . :...... :... . - . a...... :...... ' * .. ...;...:.:...... ' '..',...... *...... :: : . '..:.:...... ,.,."' .. . . .';' . ::.;.: ...... !: .!: Insulation

I Figure A-26. Pressure sensor installation. Germanium detectors, which are particularly primary pumps. Overall detection time is obviously sensitive to low level gamma detection, are often a function of primary loop sodium transport times, used in the cover gas system along with high- usually about a minute. resolution gamma spectrometers. Detection efficiency can be improved by concentrating the xenon The cover gas monitoring system can be designed and krypton isotopes in the cover gas, which is such that the fuel assembly containing failed fuel accomplished by passing the gas stream through a can be determined. Because a large breeder reactor charcoal-packed column. The time for fuel failure may comprise on the order of 300 fuel assemblies, detection by cover gas monitoring is a matter of it is important to have such a technique available to minutes. Another system often used to discover identify the offending assembly. cladding failure relies on detecting delayed neutrons that emanate from fission products circulat- One technique developed for this identifimtion is ing in the coolant stream. These neutron emissions gas tagging. Unique blends of stable xenon and are produced chiefly by two fission products, krypton gas isotopes are injected into the fission bromine-87 (56 second half-life) and iodine-137 gas plenum of each pin during final fabrication; all (25 second half-life). Both isotopes are soluble in pins within the same fuel assembly have identical sodium and enter the sodium via sodium contact on blends. A three-dimensional network of xenon-26/ exposed fuel or by fission gas expulsion into the xenon-129, krypton-78/krypton-80, and krypton- coolant. Delayed neutron detectors, usually con- 82/krypton-80 is used to yield over 100 unique gas sisting of BF3 chambers, typically are located near tags for the fuel and absorber assemblies in the

A-39 FFTEA-l49-l7 The failed assembly is identified by Sodium will burn in air matching the results of the cover gas mass spectrometer analysis with previously determined anal- e The primary loop sodium is radioactive ysis of all gas tags in the reactor, suitably corrected, for burnup and background (illustrated in Loss of substantial amounts of sodium can Figure A-27). impair cooling capacity.

Sodium Leaks and Level Measurement. The detec- The sodium level must be measured in all vessels tion of sodium leaks is important for the following containing sodium. One method to detect the pres- reasons: ence of sodium exploits the electrical conductivity

\Signal cond. 1 electronics; Imicroprocessor; [alarms It- FaiI ure Failure detection characterization subsystem subsystem Intrinsic GE detector.

controller

:over ga In 7 valve flowmeter ,/ \ ‘Sample column Absorberkol I imator mechanism Thermoelectric modules

*\ \ \ /’ \ \ / \ / \ \ \ \

Reactor vessel

Figure A-27. FFTF cover gas monitoring system. n

A-40 properties of the liquid metal. Contact type sensors within the reactor vessel, where a funciamental consist of two electrodes extended to gap at a loca- safety concern is to guarantee coolant level well tion where leaking sodium may be expected to col- above the top of the core at all times. lect. The presence of sodium shorts out the electrode gap and allows a signal to be delivered to Auxiliary Systems. In addition to the major sys- the control room. Such detectors are often placed tems discussed above in this section, niimerous under the reactor vessel and at any low point in the auxiliary systems exist, which are necessary to sup- vicinity of piping systems that contain flowing port overall plant operation. Many of these are sodium. The principal difficulties of such detectors large systems, such as heating and ventilating, but are oxidation of the electrodes and lack of assur- they are not unique to the LMFBR except for the ance that leaking sodium will actually reach the general desire to minimize the presence of water in detector. The latter is a major concern in the case of the immediate vicinity of the sodium system. This small leaks. Another type of sodium leak detector section briefly describes auxiliary systems unique checks for the presence of sodium aerosol. Atmos- to a sodium-cooled plant. pheric gas samples from the area in question can be There are numerous ways to categorize such aux- analyzed by either ionization detection or filter iliary systems. The systems described here include examination. In the ionization technique, the gas (a) inert gas, which recognizes the need for an inert stream is passed over a heated filament, which pro- atmosphere surrounding a combustible coolant, duces sodium ions. Collector electrodes will then (b) trace heating, which is necessary to keep induce an ion current if sodium is present. In the sodium in the liquid phase for low core power lev- filter technique, a replaceable submicron filter is els, and (c) sodium purification. placed in the gas stream, and the filter is periodically removed and analyzed chemically for sodium Inert Gas. An inert cover gas is a requirement for deposits. any part of an LMFBR system where a fre: liquid sodium surface can exist. Prudent design a1 so nor- Induction level probes can be used to measure the mally requires an inert atmosphere in any of the liquid level of sodium, as illustrated in Figure A-28. cells that house sodium piping systems. Whereas The induction field established by the primary coil the word inert normally implies a noble gas, the is modified by the electrical properties of the liquid essential characteristic of the atmosphere desired is coolant, thereby leading to a secondary signal that that it be chemically inert to sodium. Nitrogen satis directly proportional to the sodium level. Though isfies this requirement and is both abundant and useful for maintaining inventory records in all relatively inexpensive. Hence, it has been almost sodium repositories, this is especially important universally employed as the inert atmosphere for equipment cells. Unfortunately, it cannot be used for high-temperature application (>400°C) Signal because of the nitriding problems in the steel enclo- signal outputsignal sures. Consequently, argon has been selected as the generator j-k cover gas within the vessel, the piping systems, and the refueling transfer chambers for all major LMFBR projects to date. Helium has also been used and is still being studied as a potential alternative. The argon cover gas subsystem provides an inert atmosphere and pressure control for all liquid metal-gas interfaces. Chemical purification features include both sodium vapor and oil vapor traps. Compressors and storage facilities are necessary ingredients of this subsystem, as are pressure equalization lines to keep all cover gases at the same pressure. The purging system needs to have a high enough capacity to allow complete changes of atmosphere to accommodate maintenance operations. Because of the possibility of radioactive con- Figure A-28. Induction sodium level probe schematic. tamination, as discussed in the “Shielding”

A-4 1 section, a key feature of the argon cover gas subsys- as the RAPS system, complete with cryogenic fea- tem is the radioactive argon processing subsystem tures, but usually has substantially larger capacity. (RAPS), which removes krypton and xenon radio- One reason for the large capacity requirement is to ' .* ..isotopes. A charcoal bed using a cryogenic still is an allow pressurization of equipment cells through effective device to remove the krypton and xenon pressure testing of plant containment. isotopes from the argon stream. Surge tanks are pace Heating. Sodium melts at 98°C; hence, it useful to allow short-lived isotopes in contami- must be heated at low reactor power levels to nated argon to decay. remain in a liquid. Electrical trace heaters are normally used for such heating. A typical trace heating A nitrogen subsystem is normally incorporated assembly, illustrated in Figure A-29, consists of a to control atmospheric pressure and purity in the nickel-chromium resistant element insulated with inerted equipment cells. This is done by means of a magnesia, covered with a nickel-iron-chromium feed-and-bleed system to regulate pressure, and by alloy heater sheet, and surrounded by a large thick- fresh nitrogen purging to minimize contamination, ness of thermal insulation. Such heaters provide a using measurements of the oxygen or water-vapor heat flux around 10 to 20 kW-m2. For a large in the cells as a control signal. Nitrogen is also nor- plant, the trace heating system may consume about mally supplied for the sodium water reaction pres- 10 MW during initial start-up under cold core con- sure relief system in the steam generator, for ditions. The requirements for trace heating drop cleaning operations and for valve actuation in off appreciably when primary and secondary inerted cells. pumps are activated, from frictional heating of the pumping action. A key feature employed to remove radioactive contaminants is the cell atmosphere processing sodium Purification. The principal objective of subsystem (CAPS). It works on the same principal the sodium purification is to maintain sodium

Thermal insulation layers ( \ with staggered joints

Centerline for thermocouple fitting on 4 in. and 6 in. pipe I Centerline for \ .,AAthermocouple fitting \ \

Inner jacket 304 SST shet Outer jacket 304 SST sheel Pipe ref. Heater Heater outlet (2) Dlaces leads Heater ,,. Detail* 1 Heater orientation and jacketing c Iosu res

Figure A-29. Trace heating and insulation for 1- to 6-in. pipe. n

A-42 clean from chemical or radioactive particulate contaminants. Several trace elements from in-core Outletsodlum tC__ sodiumInlet structural materials dissolve into the flowing sodium coolant during normal operations. Table A-5 lists elements that a typical sodium purification system may be designed to monitor.

Table A-5. Contaminants monitored by a typical sodium purification system

Boron Lithium

Carbon Manganese

Cesium-134,137 Molybdenum

Chlorine Nickel

Trapping Chromium Nitrogen tempeiature

Cobalt Oxygen t Coolant gas inlet Iodine-1 3 1-135 Plutonium Figure A-30. A typical sodium cold trap.

Iron Tritium, uranium An interesting feature of a cold trap shown is the economizer. Inlet sodium must be cooled prior to entering the crystallizer, but returning or purified The main component incorporated in such a sys- sodium must be reheated to nearly the bulk coolant tem to remove impurities is the cold-trap. This temperature prior to returning to the main coolant device, connected to a by-pass line from the main system. Both functions can be performed by bring- sodium loop, removes impurities by crystallization ing the inlet sodium through a tube concentrically or precipitation at a temperature of .~150"C,signif- enclosed by an outer tube containing the coun- icantly below the main-stream sodium tempera- terflowing purified sodium. A much smaller auxil- ture. Figure A-30 illustrates a typical cold trap. iary cooling and heating system is then required to Sodium oxide crystallizes on the packing, which is allow satisfactory cold trap performancc than replaced when it begins to plug. . would be the case without the economizer.

A-43 OPERATIONAL EXPERIENCE AT LMFBWs n

Early Experimental LMFBWs . q I B.10 poison. For shutdown, in one motion the uranium section was dropped out and the B-10 section The first reactor to generate electricity was the was inserted into the region. In addition, the reac- U-235-fueled, fast Experimental Breeder Reactor 1 tor had two uranium rods that acted as control (EBR-l), which started operating in the United rods .A-9 States in 1951. This reactor, cooled by a molten sodium-potassium alloy, had an electrical output of The reactor was primarily a research facility, with 200 kilowatts. Most effort devoted to fast reactors 25 kWt at full power, so requirements for heat since that time has continued on the liquid-metal- removal were quite modest. The mercury coolant cooled concept: Much of the development on was pumped by an eddy-current type of electro- LMFBRs was carried out with early experimental magnetic pump at a rate of 0.15 L/s through the reactor plants in the United States (Clementine, core and then through a mercury-to-water heat EBR-1, Lampre, Sefor) and the Soviet Union exchanger. In the two heat exchangers, mercury (BR-1, BR-2, BR-5, BR-10). Table A-6 lists the flowed through a helical core of 2.22-mm-ID steel characteristics of these early experimental reactors. tubing inside a solid water-cooled copper cylinder.

Clementine. Clementine, the first fast reactor and A 117-cm-long 15.8-cm-OD and 15.2-cm-ID the first to use plutonium-239 as fuel, was devel- mild-steel cylinder served as the core container, the oped at the Los Alamos Scientific Laboratory. fuel cage resting on the bottom of the cylinder. Although it served primarily as a source of fast Immediately above the case, and filling the space to neutrons for experimental purposes, it also demon- the top of the pot, or vessel, was a removable reflec- strated the.feasibility of plutonium as a fast reactor tor and shield plug. The plug contained a number fuel. Operations began in November 1946 with a of layers of various materials, all in a steel cylinder power of one watt. In March 1949, the power was of 0.64-cm wall thickness (Figure A-31). A steel increased to 25 kWt. After several years of satisfac- shield, consisting of a permanent assembly of lami- tory operation, failure of at least one fuel element nations and concrete, also served as a supporting and resulting contamination of the mercury cool- and retaining wall for the reactor parts. ant led to the decision to dismantle the reactor. This was effected in 1953. Although the operating tem- In addition to the 35 plutonium fuel rods, the peratures were too low for power production, the central active core contained 20 reflector rods of reactor was used for research and it did demon- natural uranium, having the same dimensions and strate the feasibility of plutonium as a fuel for fast cladding as the fuel rods. A 15-cm-thick reflector/ power reactors. Sufficient plutonium for such a blanket of natural uranium surrounded the core. A reactor existed solely at Los Alamos during this 0.64-cm-thick aluminum jacket, containing a period. Figure A-3 1 is a cross-sectional view of the water-cooling tube, removed the heat generated in reactor, a rectangular block-like structure. the uranium blanket. The core and blanket were surrounded by a 15.2-cm steel reflector and a 10-cm The 15-cm diameter core contained 55 fuel and thickness of lead shielding. The top shielding was blanket elements. The fuel rods, of delta-phase plu- made of a series of blocks that could be removed to tonium, were 1.64 cm in diameter and 14 cm long, give access to the reactor.A-18 Since an important were clad with a type 1020 steel, 0.5 m thick, and purpose of this reactor was to provide fast neutrons assembled in a vertical for experimental purposes, a number of holes were provided for radiation work. Four horizontal holes Control of the reactor was effected by several ran completely through the reactor. In addition, components. A reflector, consisting of a large three reentrant horizontal holes and 10 vertical block of uranium immediately below the active holes were provided. region, could be raised into position and increase the reactivity, and could be quickly dropped out to Following attainment of criticality in November reduce it. llvo safety rods, located in positions 1946, the reactor was operated as a critical assem- immediately outside the central region, were com- bly; considerable information was obtained con- posed of a section of uranium in a section of cerning the physics of fast reactors. During n

A-44 Table A-6. Early experimental LMFBRs

Reactor (Country)

CLEMENTINE EBR-1 BR-1 BR-2 BR-5 LAMPRE SEFOR BR-IO (USA) (USA) (USSR) (USSR) (USSR) (USA) (USA) - (USSR)

General

Date critical 1946 1951 1958 1961 1969 1973 Date full power 1949 1951 1959 1961 1971 - Electrical power, MW - 0.2 - - - - Thermal power, MW 0.025 1.2 5 1 20 10

Core parameters

Core volume, L 2.5 5.9 17.2 3.1 566 - Core height, rn 0.14 0.22 0.28 0.15 0.93 - Core diameter, m 0.15 0.18 0.28 0.16 0.88 - Core volume fractions - -

%Fuel - 48 -52 43 %Sodium - 15** -34 30 %Other - 37 -14 27

FI~X,1015 n/cm’.s 0.004 0.11 1 .o - 0.6 20

Power denisty, kW/L - 170 35 - A .F*S - 0.04 0.6 - Peff - 0.0068 0.0032 - Doppler constant, Td/k.dT - - 0.008 -

Fuel

Fuel type Pu Enr. U PuO2,UC -a Fuel pellet forms R S S S

Reflector/blanket

Axial material - U Ni - Radial material U U Ni U

Control

Material l0B+U U Ni Ni + SS Ni - No. assemblies 4b 12 2+2c 4d 10 -

Reactor vessel

Height, m 1.2 0.7 Diameter, m 0.2 0.4 Thickness, m 0.002 0.01 .

Containment structures

Configuration I I D I Materials - 2 - -

Primary heat transport

Coolant Hg NaK None Hg Na Na Na Na Cover gas - Ar - - Ar - Ar -

A-45 Table A-6. (continued1

Reactor (Country)

CLEMENTINE EBR-1 BR-1 BR-2 BR-5 LAMPRE SEFOR BR-10 (USA) (USA) (USSR) (USSR) (USSR) (USA) (USA) _(USSR)

- Loop Loop Loop Loop Type Loop Loop - - - 2 2 le 2 No. loops 1 1 - C E E Pump type E E - E - - 36 - - 60 7 Total flow, kg/s 2 - Inlet core temp.,"C 38 230 . - - 430 450 370 - Outlet core temp.,"C 121 322 - 60 500 560 430 - - Max fuel temp.,"(= 135 477 - - 1300 870

a. Molten plutonium alloy.

b. Radial uranium reflector also used.

c. 2 Control rods + 2 cylindrical reflectors.

d. 4 Control rods + annular reflector.

e. Auxiliary loop also used.

aCwret.

~omnPlastic

'saSteel

Led Brick

0Tut.dloy

Bismuth

bonito

0Thorium

Concrete Md Led Shot

Figure A-31. Cross section to the CLEMENTINE reactor shield

A-46

I low-power operation of the reactor from February Control 1947 to January 1949, measurements were made of the critical mass, effectiveness of reactor control, temperature coefficient, neutron spectrum, and general behavior. The information obtained was valuable in establishing the feasibility of fast reactor operations, including the demonstration of control by delayed ne~trons.~-l~

From March 1949 to December 1952, when the rupture of a plutonium fuel rod was noted, the reactor was used for fast-neutron radiation research, as well as for a continuing program of reactor physics research. As a result of the fuel rod rupture, plutonium was released into the mercury Electromagnetic ’ coolant circuit, and since the primary objectives of pump the experiment had by then been achieved, the reactor was dismantled.A-20 Figure A-32. EBR-I heat transport systern.

EBR-1. The Argonne National Laboratory Experi- average heat flux of 690 W/m2. Although these mental Breeder Reactor-1 (EBR-1) is located at the values have been exceeded in more recently Idaho National Engineering Laboratory, Idaho. It designed fast reactors, EBR-1 may be considered was designed to demonstrate fast reactor breeding the first high-power density reactor. Initia 1 plans and to prove the use of liquid-metal coolants for called for a conservative design that would permit power production. As a model, a power level of operation at power levels greater than 1 MW, but 1 MW thermal was chosen. Electrical power gener- heat generation in the outer blanket proved greater ated about 200 kW(e).A-21 than anticipated, and limited operating pow,~r.~-~~

EBR-1 was based on concepts proposed in 1945. Core Arrangement. Figures A-33 and A-34 show The research and development program was the Mark-I1 core arrangement. The inner fuel rods, approved early in 1945, and the Manhattan Engi- 1.14-cm OD, were separated without spacer ribs by neer District approved construction in November 1.7 cm (the Mark-1 core used spacer ribs). 1945. Design and construction occupied the years Enriched fuel was used in the middle of each rod, from 1948 to 1951. Criticality with the Mark-1 core and natural uranium was used at the top arid bot- was reached in August 1951, and electricity was tom to form an axial blanket. An inner blanket con- generated on December 22, 1951. The Mark-II sisted of natural-uranium slugs, 2.58 cm in core was installed in 1954. A series of kinetic exper- diameter and 51.44 cm long, jacketed in 0.0.56-cm- iments led to its meltdown on November 20, thick stainless steel. The core itself and the inner 1955.A-22-26 blanket were cooled by circulating sodium- potassium (NaK) alloy. The Mark-I and Mark-11 cores were similar, with the exception of some changes in spacing ribs and An air-cooled outer blanket is located outside the composition, as described in References A-17 reactor tank. It consists of 84 keystone-shaped through A-31. The Mark-I11 core was designed pri- natural-uranium bricks, each weighing 45 .5 kg, marily to investigate the stability of the fast reactor. clad with stainless steel 0.051 cm thick. This sec- The Mark-IV core, loaded in 1962, used plutonium tion is movable and contains the control rods Mov- as a fuel and provided a facility in which the general ing parts are kept outside the liquid metal. Air operating characteristics of a solid-plutonium cooling of the blanket proved to be the limitation fueled reactor could be in~estigated.*-~l-~~ on the operating power available for the reactor. Surrounding the external blanket is a graphite Reactor features. Design of EBR-1 was a pioneer- reflector 48.3 cm thick, followed by 2.74 m of con- ing effort; much of the present nuclear reactor tech- crete shielding. Six experimental beam holes pierce nology had not yet developed. Figure A-32 is a the concrete shield and graphite reflector. A ther- drawing of the reactor installation. The Mark-I mal column and “rabbit hole” also provide facili- core had a power density of 170 kWt/liter, and an ties for experiments.

A-47 -NOTE I, EXTERNAL BLANKET SEGMENT TYPICAL FOR CONTROL OR SAFETY ROD 2. I EXTERNAL CONTROL ROO IN EACH OF 4 SEGMENTS AN0 I EXTERNAL,SAFETY ROD IN EACH OF 8 SEGMENTS

IN ALUMINUM ----

Figure A-33. EBR-I Horizontal cross section at midpoint of MARK-I1 core.

Fuel. Most of the rods for the Mark-I core were Control and Instrumentation. There were 12 con- loaded with four 0.923-cm-diameter slugs, each of trol rods, each 5.08 cm in diameter, made of natu- which was 4.76 cm in length. Some of the rods, ral uranium with a jacket of stainless steel. These however, contained 6.35-cm-long, 0.975-diameter moved vertically in the outer blanket brick. Eight slugs. Below the fuel section, an 11.4-cm-long of the control rods normally were used as safety natural-uranium slug served as a lower blanket, and rods. Their time of travel out of the blanket was above the fuel section a 20-cm-long natural- short: 0.085 seconds to initiate motion, 0.29 sec- uranium slug was loaded as the top blanket. onds to reach 40.6 cm. The remaining four rods normally were used as regulating control rods and During operation of the first core, experiments could be positioned with considerable accuracy. showed that an alloy fuel composed of 2 wt% U-Zr Their maximum speed was 1.63 cm/s. The entire was more resistant than beta-quenched uranium to outer blanket was mounted on a hydraulically- irradiated growth and was free of the irradiation- driven elevator. The main platform of the elevator induced surface-roughening characteristic of beta- carried a shield section on which the outer blanket quenched uranium. Early in 1954, a Mark-I1 core rested. Arrangement permited accurate location of was installed with a 2 wt% U-Zr alloy in both the the outer blanket around the reactor. For shutdown fuel and blanket slugs. The fuel slugs were 10.8 cm the reactor blanket, the fuel plug on which it rests, long by 0.975 cm in diameter, with a lower blanket and the elevator could be dropped quickly. slug of 10.8 cm and an upper blanket 20 cm long. As in the Mark-I core, the annulus between the slug For Mark-IV, about 45% of the safety circuits and fuel tube was filled with NaK as a heat-transfer provided signals indicating abnormal operations, bond. five provided signals indicating danger (designed to

A-48 I\ Fuel element

'uc t ure rods

Reactor tank

Control rod

Safety b lock

Figure A-34. EBR-I Mark-I1 reactor-cutaway view. ~. scram the reactor after 2 minutes without correc- Heat Transport System. The primary Nali flow tion), and 16 scrammed the reactor immediately. through the reactor, normally 18.4 L/s, was from an The immediate scrams were designed to react when elevated constant-level tank, shown in Figure A-32. any one of the following parameters left its speci- The flow proceeded down through the inner tilanket, fied range: reactor period, fuel and blanket temper- up through the core, and out of the reactor. The cool- ature, coolant flow rate, inlet and outlet coolant ant then flowed through the intermediate heat temperature, elevator hydraulic pressure, reactor exchanger, returned to a receiving tank, and was then neutron-flux level, or position of any of the four pumped continuously by an electromagnetic pump to NaK valves. All safety circuits were duplicated to the constant-head tank. increase reliability. In addition, checklists were used periodically by the operators, and frequent In EBR-1, the NaK-to-NaK intermediate heat detailed inspections of all equipment and safety exchangers were shell and tube design, with pri- devices were mary flow passing through the tubes. The steam generator was divided into an economizer, a boiler, Operating controls contained a total reactivity and a superheater. NaK passage through these units worth of about 1.28%. Safety rods and safety plugs was countercurrent to the flow of water and steam. could remove 0.24% reactivity in 0.38 s, and the Heat transfer tubes in each component were similar external blanket could remove 5.25% reactivity at and consisted of a composite assembly of' inner an elevated speed of 66 cm/s. The above tempera- nickel, intermediate copper, and outer nickel tubes. ture coefficient for Mark I was 3.5 x Ak/"C The tubes were assembled by a mechanical drawing from 38 to 200°C. Filling with the coolant added process, together with a thermal diffusion bonding reactivity approximately equal to that added by process, which resulted in good heat transfer 2 kg of U-235. Loss of coolant reduced reactivity. between the tubes. Total wall thickness of the tubes Separate instrumentation was provided for start- was 0.794 cm, of' which 0.476 cm was nickel. An up, for safety, and for a steady operation, and all outer stainless steel tube made up the shell of the safety circuits were duplicated. heat exchanger, and a bellows was used to allow for

A-49 differential thermal expansion. Thus, each heat shield by absorbing gamma rays. In the axial direc- exchanger was a single-tube-in-a-shell type, for tion above the core, layers of steel and NaK were n NaK flow was on the shell side, and water or steam followed by a 33-cm-thick steel plate, a 5.7-cm was in the tube. plate, a 61-cm thickness of concrete, and finally, 20.3-cm thickness of laminated masonite and iron. A forced-air circulation falling-film boiler was The bottom shield consisted of a 1.52-m-thick layer used to limit the quantity of water in the system and of ordinary concrete. Since the elevator room was to increase the heat-transfer rate. Heat exchangers the only area below the reactor, and since it was a were vertical. A baffle was used to establish a water limited access area, the bottom shield design was film at the upper end of the internal tube on its not considered critical. Additional details of the inner surface. The film ran to the bottom where EBR-1 shield system are given in Reference 19. excess water and generated steam were piped into a drum. Steam was led through a separator in the ,4t the time EBR-1 was designed, much shield drum out to the superheater, which consisted of design theory had not been developed. One of the horizontal heat exchangers with NaK in the shell purposes of the reactor was to improve the concept side. The economizer was horizontal and served to of fast reactor physics and permit experimental heat the feedwater from the deaerating tank to measurements of neutron source distribution and steam temperatures before injection into the boiler other parameters. Flexibility was incorporated in drum. the design to permit addition of more shield thickness if necessary. Since the EBR-1 was of modest Reactor VesselandShielding. The reactor vessel, or power and was located in an unpopulated area, a tank, was double-walled and extended through the containment type building was not used. reactor shield. The section of the reactor vessel surrounding the reactor core had an inside diameter of Operating Experience. When the Mark I reactor 90.3 cm and a length of 71.1 crn. Above this small was started up in 1951, radiation levels around the section, the vessel increased in diameter and was reactor were found higher than anticipated. An filled with shielding material, mostly steel. The additional 76.2 cm of concrete shielding was whole reactor vessel rested on the shoulder formed installed, and operation was resumed. In June by the change in diameter; thus, the reactor core 1952, a NaK leak was discovered in the primary itself projected below the point of support as a heat exchanger, and the reactor was shut down for smooth cylinder. repair. During the shutdown, 16 fuel elements were removed for examination and were replaced by new The small diameter part of the reactor tank con- elements. More than 1500 MWt hours of operation sisted of a stainless-steel vessel with a 0.794-cm- had been accumulated by April 15, 1953. thick wall, made by deep drawing. It was surrounded by a second tank made of Inconel, 0.16 cm thick, which fit snugly on the ribs formed During the four-year period, EBR-1 gave essen- in the Inconel. The upper portion of the reactor tially trouble-free operation. Operation of EBR-1 vessel also was double-walled. The gas space demonstrated, among other things, that breeding between the two walls provided some thermal insu- was technically achievable with a measured conver- lation and provided a method for testing vessel sion ratio of 1.01 k 0.05, that the use of liquid integrity. In the event that the inner vessel should metal coolant (sodium-potassium alloy) was com- develop a leak, the outer vessel would prevent compatible with breeding economy, and that the system plete loss of sodium. was metallurgically and mechanically feasible. Breeding, based on U-235 fuel, was demonstrated A 46.7-cm-thick 89-cm-high section of graphite in the first cores. Breeding in a plutonium cycle was surrounded the outer blanket in the radial direc- demonstrated with operation of the Mark-IV tion. Its cooling-air gap is shown in Figure A-33. core.37 The EBR-1 operation and theoretical deter- This graphite layer, in addition to serving as a mod- minations also showed that neutron behavior below erator for fast neutrons for shielding purposes, prompt critical is the same in both fast and thermal acted as a reflector to improve the breeding in the reactors. outer blanket. Although ordinary concrete was the basic shielding material, a 10.2-cm layer of iron was Under normal operating conditions, the reactor provided between the graphite and the 2.59-m- was very stable and did not exhibit a prompt, posi- thick concrete cylinder; this served as a thermal tive temperature coefficient or a resonance. Under n

A-50 purposely imposed and drastically abnormal oper- nor any other person was injured in any way. A ating conditions, anomalies were observed: reso- negligible amount of radioactive material reached nance consisting of oscillations in power levels the atmosphere through temporary therm'xouple appeared during experiments in which the coolant wire seals. Evacuation steps were precautionary. flow rate was drastically reduced; a prompt positive Operating personnel returned to the reactor build- temperature coefficient appeared during startups ing after a minor amount of surface contamination undertaken with reduced coolant flow. But even was removed. The core assembly was removed from under conditions where the net positive coefficient the reactor by use of a temporary hot c,ell and appeared, the reactor could be operated safely. shipped to Argonne National Laboratory fc'r exam- Oscillator tests successfully demonstrated the pres- ination and disassembly. Observations during dis- ence of instability. assembly and subsequent simulated meltdown experiments indicate that the porous structure In November 1955, the Mark-I1 core of EBR-1 formed in the core could have resulted from the partially melted during the last of a series of experi- vaporization enirained NaK.A-38 ments designed to study its behavior when put on positive power periods with reduced or zero coolant Mark-/// Program. After the Mark-I1 incident, a flow. The accident occurred under extremely new core was designed to further study thl: previ- abnormal operating conditions, purposely ously observed instabilities, with the belief that imposed on the reactor for the experiment, and rec- they could be eliminated by changes in the niechan- ognized to involve a risk of fuel melting. Two of the ical design. Other objectives included an investiga- normally operative safety mechanisms, the flow tion of channel transient characteristics, such as interlock (which automatically shuts down the measurement of transfer functions, reactivi1.y feed- reactor if substantially full coolant flow is not back effects, arid development of mathematical maintained) and the period scram meter interlock models to describe the observed results. (which automatically shuts down the reactor if the period becomes too short), were purposely discon- Figure A-35 shows a typical fuel rod used on the nected. Coolant flow was stopped completely. A Mark-Ill core. Cladding consisted of 0.051-cm of certain fixed amount of reactivity was put into the zircaloy metallurgically bonded to the fuc:l by a reactor with the control rod, and the reactor was coextrusion process, rather than by the arrange- started up on a short enough period so that a severe ment used for lvlark-I and Mark-I1 where NaK temperature differential would be established in the served as a heat transfer bond between the loose fuel slugs. The prompt positive temperature coeffi- fitting slugs and the stainless steel can. On each cient previously observed appeared, and, as the rod, three 0.117-cm wires were spot welded to the power increased, the reactivity increased, thus fur- fuel cladding at 0.64-cm intervals, which served to ther shortening the period. It was planned to scram stabilize the rod. At full loading the core contained the reactor when the period reached 0.27 s and the 252 rods, or 60 1:g of enriched uranium. The cen- temperature of the fuel 500°C. When the period tral core section of each rod consisted of three slugs reached 1 second, the operator mistakenly acti- welded end-to-end; in the lower and upper blanket vated the slow-acting motor-driven control rods section, each rod consisted of a single slug of instead of the faster-acting scram rods. By the time 2 wt% U-Zr alloy, 9.05 cm long and 19.7 cni long, the scram was initiated, the period had reached respectively. A triangular tip at the bottom of each 0.3 seconds. The temperature overshot, so that the rod was provided to simplify insertion in the lower uranium became heated above 720°C, roughly the tube sheet during the loading operation. Eilanket temperature in which the uranium-iron eutectic rods were of identical design, consisting of a single formed. The center of the core melted, forming the 48.7-cm-long section natural-uranium zirconium . eutectic. After the manually operated scram button alloy. was pressed (in less than 2 s), the reactor shut down and the meltdown stopped. The automatic power- Fuel rods subassemblies, shown in Figure A-36, limiting circuits also operated. contained 37 individual rods supported in a grid at the bottom of each subassembly. A nozzle admitted As a result of the accident, melting occurred in the NaK to it. One central position was occupied by 40 to 50% of the EBR-1 core. No explosive force an expandable tightening rod that forced the rods developed. None of the remainder of the reactor, outward against the hex wall and limited radial including the inner blanket and the reactor vessel, movement of the fuel rods. Seven fuel assemblies was damaged. And neither the operating personnel composed the core assembly, as shown in

A-5 1 n

. i \-STAINLESS-STLEL HANDLE --

24 (87 IN a?

LA j- 8.187 IN STAINLESS-STEEL HANDLE LIRCALOY-2 CLAWINQ 10046 IN. WIRE DIA. 0020 IN. THICK1

Zr OkO . DIA.

SECTION A-A

Figure A-35. EBR-I Mark-I11 fuel rod.

SECTION A-A

SECTION B-B

Figure A-36. EBR-I Mark-I11 fuel and blanket assemblies. n

A-52 Tie rod Downcomer

subassemblies -J subassemblies

Figure A-37. EBR-I Mark-I11 core cross section. Figure A-37. Blanket subassemblies, each contain- located between the inner edge of the seal phte and ing 37 rods, had a different design for coolant flow. the blanket assemblies restricted the leakage. Twelve of the blanket subassemblies surrounded the core. Double-wedge clamps were located on each of The coolant, in series flow, entered the annular the six flaps at the outer periphery of the assemblies inlet plenum located immediately below the seal at the core center line, and served to force the outer plate and flowed into the outer ring of the twelve assemblies inward against the center assembly. A blanket assemblies. At the bottom of the blanket second set of six shoe-type clamps mounted along assemblies, the flow was reversed 180 dN:grees, the inner edge of the seal plate limited the by-pass directed up three to seven fuel assemblies, through leakage rate and served to lock fuel and blanket the outlet holes at the top, and then radially out assemblies into a rigid array. Also shown in through the perforated portion of the tdanket Figure A-37 are the twelve downcomers through assemblies into the outlet plenum. The coolant, in which the coolant passed in parallel flow. The six parallel flow, flowed into a lower annular plenum tie-rods held rigid the lower structure, the located immediately above the mounting plate. antimony-beryllium source, and the oscillator rod Here the coolant was distributed to the twelve and thimble. downcomers through which it flowed to the lower plenum. Upward flow through fuel and hlanket assemblies was partitioned by a series of throttle valves. The actual flow through the core, both in series and parallel, was less than the flow indicated through the metering of a primary inlet. Of a nomi- plenum restricted by-pass leakage occurring nal metered flow of 18.3 L/s, approximately 16% between the tank and the outer edge of the seal was bypassed as leakage and for seal plate cooling plate. Restriction was accomplished by a system of in series flow. The remaining 84% passed through two expandable Inconel seal rings. Seal plate shoes the blanket and core. For a metered flow cif 11.8

I A-53 r/,re-clamp actuator

~ Seal-plate clamp act1Jator

Inlet-valve actuator

Throttle valve

Reactor tank Inner blanket-rod

Outer blanket

Lower shield

Figure A-38. EBR-I Mark-I11 inner tank assembly.

A-54 L/s in parallel flow, approximately the same frac- behaviors of the partially sheared core resulted in tion, 84070, passed through the core; the remaining the following values for the respective prompt- 16% passed through the blanket. Coolant from the negative, rod bowing, and delayed structural power blanket outlet cooled the seal plate. coefficient components:

The effects of rod bowing were explored by sys- -2.21 x 10-6 tematically removing part of the stiffening ribs. Although the rods were rigidly fixed at the upper +0.543 x and lower end, a radial motion in the fuel section accompanied differential expansion during a power -0.873 x Ak/k-kW. increase and resulted in a bow directed at the center high-flux region. In 14 tests involving various types These values correlate in a curious manner with of changes, transfer function measurements those empirically deduced for the prompt-negative showed a progressively larger positive effect con- and delayed structural power coefficient compo- sistent with the removal of the ribs. The amplitude nents for the fully ribbed core, namely, -2.21 x of the power feedback decreased, and the phase lag and -0.330 x Ak/k-kW. increased as ribs were removed. The magnitude of the positive power coefficient It was concluded that the fully-ribbed and rigid component is equal to the increase and the magni- Mark-I11 design was completely stable under normal tude of the structural component. Apparently rib operating conditions. The ribs provided strong radial shearing introduces two feedback processes, rod coupling between individual fuel rods and between bowing, which is prompt and positive, and one of the fuel rods and the hexagonal can, and therefore an unspecified nature, which is negative and they provided a large radial contribution to the power extremely delayed in time. coefficient. The positive feedback effect from the , inward bowing of the rods during power increases was Because the model predicts a general shape of the likewise eliminated. Testing indicated that the core experimentally measured feedback and comes could be brought into a resonant condition at an within a reasonable margin of explaining the feed- extrapolated power above lo00 MWt, well above the back magnitude, it was concluded that the struc- design value of 1.2 MWt. tural member responsible for the delayed negative power coefficient was the lower shield plate. A Power coefficient nonlinearities were observed, credible interpretation of the feedback would which complicated the interpretation of test data in involve the concept of a delayed feedback that terms of a single model that could be applied over probably originated at some structural member wide ranges of power, flow, and temperature condi- located downstream of the core. A careful scrutiny tions. These nonlinearities apparently resulted of all downstream structural members supple- from the power and temperature sensitive clear- mented by attenuation concepts narrows the sus- ances existing between rods, between rods and pect region to that included between the lower hexes, and between hexes, which are complicated shield and seal plates. The fact that thermally by differences in the coefficient of expansion of the induced ligamenta1 motion has actually been materials involved. Furthermore, the power coeffi- observed in such a plate, and that the maximum cient itself was found to be sensitive to the tempera- temperature differential across the shield plate ture inlet cooling. occurs close to the natural resident frequency of other reactors, strongly supports this conclusion. At one-third flow, the reactor would reach reso- The strong increase in the negative structur: term nant instability at a power of about 10 MWt. At does not affect conclusions regarding the resonant full coolant flow, with the ribs sheared, feedback stability at the fully or partially sheared cores. The measurements indicate a probable resonant insta- process acted so slowly-the reactor oscillated at bility at a power level of 1 l MWt. frequencies as low as 0.02 cycleds- that it cannot sense the feedback. It has also been shown that rib shearing results in an unexpected and unexplained increase in the The Mark-111 core, even with ribs partially magnitude of the delayed structural power coeffi- sheared, was found to be much more stable than the cient component. An empirical fit of feedback data Mark 11. In addition to elimination of the inward between models describing the dynamic and static rod bowing found in Mark-I, the Mark-I11 (design

A-55 resulted in a negative prompt power coefficient runs at low and moderate power indicated that the contributed by both axial and radial expansion of reactor was quite stable, the system was gradually the rods. In Mark-11, the negative coefficient asso- brought up to power, 900 kW thermal, during ciated with axial fuel expansion and coolant expul: April 1963. Although operation was very satisfac- sion was insufficiently large to cancel the pkompt tory, some loss of reactivity was noted. component from rod bowing. It was therefore concluded that the instabilities noted in Mark I1 were EBR-1 was shut down and secured on completely the result of the design peculiarities and December 3, 1963. Total power produced by the not an inherent feature of the reactor concept itself. Mark 1V loading was 577 MW/hrs. Eleven fuel After conclusion of the stability test in 1960, the rods were removed from the reactor for return to reactor was used on a part-time basis to study vari- the metallurgical division of Argonne National ous loading arrangements, radiation effects of dif- Laboratory for examination of the irradiated pluto- ferent types of fuel elements, and fuel isotopic nium fuel plugs. changes. The reactor was secured subcritically by Mark-IVCore. The full potential of breeding with 1.75% Ak/k with the controls in their most reac- a fast spectrum is realized only when plutonium, tive position, and the reactor temperature at 30°C. rather than U-235, is used as the fuel in the core. Surveillance of the reactor blanket gas was main- The EBR-1, Mark-IV core, loaded in 1962, was the tained during this indefinite, extended shutdown first use in this country of plutonium for the full period. core of a fast breeder power reactor. Operation of this reactor facility with plutonium loading had as its objectives the determination of general operat- LAMPRE-I ing characteristics, the measurement of breeding gain, and the determination of radiation effects Molten Plutonium Core Concept. Interest in the and nuclear parameters. development of plutonium fuels for fast breeder reactor applications continued at the Los Alamos An alloy of plutonium with 1.25 wt% aluminum Scientific Laboratory after the dismantlement of was used in the core. Slugs of the material, 0.59 cm Clementine (see previous section on Clementine, in diameter and 5.4 cm long, are contained in p. A-44. Parallel to this effort has been the develop- zircaloy-fuel tubes, 0.76 in. OD with a 0.053-cm ment of homogeneous reactors wherein the fissile wall thickness. Three full-length ribs served to cen- material was dissolved in an aqueous system. Reac- ter the slug and the jacket tube. Each fuel rod con- tors of this type include the water-boiler series, as tained a lower depleted uranium blanket slug well as LAMPRE-I and LAMPRE-11. The resulting 8.95 cm long, four plutonium alloy fuel slugs total- specialization in plutonium-fueled fast breeder reac- ing 21.6 cm, and a single 19.67-cm-long upper tors on the one hand and the homogeneous concept blanket depleted uranium slug. A 0.0318-cm NaK on the other logically led to the concept of using layer bonds the cladding to the fuel. A number of plutonium in the molten state as fuel for a fast rods had a 0.2-cm-diameter zircaloy thermocouple breeder system. An early reactor design, for exam- tube attached. Blanket rods were similar to those ple, called for the plutonium fuel to be obtained in a used in the fuel section; depleted uranium slugs cylindrical vessel through which tubes carrying were substituted for the plutonium alloy. sodium coolant would flow in a typical calandria arrangement. However, the need for additional At shutdown, the Mark-I11 core had operated for information regarding the behavior of container and 3,220 MW/hr. The core was completely unloaded fuel material led to the decision to first build a mol- by November 8, 1962, and the Mark-IV loading ten plutonium fueled reactor; within such a reactor, went critical on November 27, 1962, with 27 kg of the fuel would be contained in cylindrical capsules plutonium-239 (327 fuel rods). Initial loading con- with the sodium coolant flowing outside. This sisted of 60 fuel rods in the central assembly. Ten arrangement, used in LAMPRE-I, could be readily subsequent loadings, four of 42 rods each, two of adapted to the testing of a variety of fuel and con- 24 each, two of 18 each, one of 8, and a final one tainer material combinations. A cell in an existing of 7 rods were made to reach criticality. Foil irradi- building, which had previously been used for the ation runs at low power were carried out in early LAMPRE-I reactor, was available for LAMPRE. 1963 to determine constants necessary for the cal- The limitations of this location fixed the reactor culation of breeding ratios. After transfer function power at 1 MW(t).A-30-51

A-56 Reactor Features. A plain view of the reactor 15.2 cm high. Flow continued through a locator installation is shown in Figure A-39 and an elevated plate assembly, then finally passed the fuel cap- view in Figure A-40. Component design was based sules, through a top reflector region, and into the on the need for installation in the existing cell facili- outlet plenum. ties. No secondary sodium coolant is used. The The core arrangement contained several safety heat generated is exchanged to air and is exhausted features. Double-wall construction of the reactor up a stack. Although thermal performance specifi- vessel, with no pipes entering the lower part, pre- cations are high, specific power and power density vented accidental drainage of coolant. A catch-pot were not intended to represent optimum values for and diluent plug were designed to contain any fuel, a larger power reactor system. in a noncritical geometry, in the event of a leak in the core. The Armco@ iron diluent plug would dis- Core Arrangement. The core consisted of approxi- solve in molten fuel to form an alloy with a higher mately 140 fuel capsules filled with plutonium iron melting point. As the solution continuzd, the alloy surrounded by approximately 60 stainless resulting alloy would solidify. steel reflector pins of similar design. A cross- section of the core region is shown in Figure A-41, Fuel Elements. A single fuel capsule or “pin, ” as and a vertical section of the core is shown in shown in Figure A-43, was used in each complete Figure A-42. An annual, moveable stainless steel element shown in Figure A-44. The components of reflector was contained in an inner vessel outside the capsule assembly shown are the thimble, fuel the core. It was 50.8 cm in OD by 27.3 cm ID by slug, closure, plug, and adapter. The capsule fuel 40.6 cm long. Final control was obtained by mov- thimble was constructed of various tantalum types, ing four control rods, each consisting of a nickel depending upon the desired materials test. A typi- cylinder 9.65 cm in diameter and 9.46 cm long, cal thimble, fabricated from tantalum of 0.1 wt% which move vertically in a stainless steel reflector. tungsten, was 1.08 cm OD and 20.1 cm in overall length, with an ID tapered from 0.955 cm at the top Coolant sodium flowed down to a 0.95-cm annu- to 0.919 cm at the cone end. Solid plutonium iron lus between the vessel and the flow divider. The alloy fuel slugs were machined immediately before coolant stream reversed in a plenum at the bottom assembly into the thimbles. These fuel s1u.g were of the flow divider. The sodium then flowed 0.909 cm in diameter, and lengths varied depending through a bottom reflector consisting of an on the fuel weight required. However, the lengths Armco@ iron cylinder 17.5 cm in diameter by averaged 16.1 cm.

Figure A-39, LAMPRE-I plan view.

A-57 Figure A-40. LAMPRE-I elevation.

Figure A-41. LAMPRE-I horizontal cross section.

A-58 --

-S 'odiun I inlet

-Borated- graphite cylinder

-Shim m T I c L.-

Figure A-42. LAMPRE-I vertical cross section.

A-59 Tantalum NRC temescal Closure plug elec tron-beam vacuum and adapter melted Tantalum NRC Length to be determined commercial grade rod I- by weight specification -1 -1 -1 0358 IN DIA f Pu-Fe alloy slug

Figure A-43. LAMPRE-I fuel capsule and slug.

Top locking sect. 0.483 across 0.500 across flats

Section A-A

Figure A-44. LAMPRE-I fuel subassembly. The remainder of the assembly, shown in included a number of accessory components: flow- Figure A-44, included a 2.29-111 shielding section meters, a heating transformer for raising the tem- and a so-called capsule handle. The capsule handle perature of the flowing sodium to the melting point was constructed of Armco@ 17-4 stainless steel of the fuel, three getter hot traps, and a fill and and was used to insert and withdraw the fuel cap- dump tank system. sules from the core. It was also used to maintain the Shielding. The shielding requirements for this radial core configuration. installation were somewhat unusual since the reactor was installed in existing facilities adjacent to the Cooling System. The cooling system consisted of control room. The borated graphite shield, 1.07 m two parallel Callery A-C electromagnetic conduc- thick, surrounded the reactor vessel. A 20.3-cm- tion pumps. The pumps, each rated at 63 L/s at thick lead curtain and a 1.68-m normal concrete 1,190 n/m2 head, were used for circulating sodium wall shielded the control room. Control room radi- to the reactor. Heat was removed directly from this ation levels during initial operations were found to primary circulating system to air by a finned sec- be excessive. The installation of improved shielding tion heat exchanger, which exhausted up a stack. in the concrete wall penetrations and a supplemen- The coolant loop, constructed of a 5-cm and tary lead shield reduced the levels to below toler- 7.62-cm Schedule 40 Type 316 stainless steel pipe, ance values.

A-60 . .-......

A vertical cross-section of the shielding is shown and containers, and investigation of operating in Figure A-45. The shield below the core was based problems that might be unique to the fluid nature i on a design specification of 2-mrem/h dose rate in of the fuel. The reactor started in 1961 with dry the exclusion area in the shadow of the shield. In critical and low-power operations. It was brought addition, the shield consisted of a 40.6-cm-thick up to its design power of 1 MW in early 1961. In iron bottom reflector in a 2.13-m-long floor shield April 1962, the reactor was reloaded with a h4ark-I1 plug filled with lead shot in the bottom 30 cm, and core. After capsule failure in September 1962, magnetite aggregate in the remainder. which permitted 75 g of plutonium iron fuel to enter the coolant, the power level was temporarily The fuel-capsule handles alone provided approx- limited to 500 kW. imately 2.7 m of iron shielding directly above the core. A ceiling shield plug of heavy concrete and steel shot provided the access through the 1.68-m Stabilityand Control. Although bubbles of fission- concrete ceiling. During initial low-power opera- produced gases tend to rise to the surface of the tion, it was necessary to add an additional lami- fuel, some separation of large portions of the fuel nated iron masonite shield on top of the shield plug regions apparently result in reactivity changes with and a concrete shield around and on top of the fuel time, as shown in Figure A-46. In the Mark-I load- transfer area. ing, solid additives of carbon or plutonium carbide used to inhibit corrosive attack on a tantalum cap- Operation. The LAMPRE-I reactor was designed sule were believed to interfere with the release of the as a facility for the study of molten plutonium fuels accumulated gas, accentuating this reactivity loss. Sod iu rn eq u iprnen t room Ceiling shield plug \ 6 in. Steel 7 14-in. steel cover

036u Feet

Figure A-45. LAMPRE-I shield, vertical cross shielding.

A-6 1 0 2 4 6 81012 Time (min)

0 10 20 30 40 50 60 70 80 90 100110 120 Integrated power [MW(T)] Figure A-47. LAMPRE-I core power and inlet and outlet temperature changes from 45 kW(t) vs. time. Figure A-46. Reactivity loss during LAMPRE-I operation.

Although approximately $10 of reactivity was lost in the Mark-I loading, there was never any indication of instability caused by the froth or low density portion of the fuel. In the Mark-I1 loading, the disengagement of evolved gases was cleaner, with a smaller reactivity loss; this resulted as a function of integrated power, as shown in Figure A-46. During operation the dynamic characteristics of the reactor -8-6wi I I111111 I I1111 were evaluated. Temperature coefficients, power IO' 1 10 102 Frequency (radiansls) coefficients, and reactor transfer functions were measured. Figures A-47 and A-48 show typical results. Substantial negative temperature coeffi- Figure A-48. LAMPRE-I transfer functions for various cients provide a high degree of operational stability. power levels. operating temperatures. Some test capsules con- Fuel Test Program. LAMPRE-I fuel consisted of taining the former alloy have been irradiated in the 24 g of plutonium alloy, 90 at.% plutonium, LAMPRE core. 10 at.% iron. The rather unusual density characi teristics of this alloy (Figure A-49) can lead to a Following fuel leakage from several elements, the plug during solidification because the lower density LAMPRE-I was shut down and unloaded. Deci- solid material tends to float to the top of the sion not to reload the reactor was based on a find- remaining fluid portion. In the initial loading, pres- ing that little, if any, new information could accrue ence of a few hundred ppm of carbon and a pluto- from continued operation of the reactor. The maxi- nium fuel alloy was believed to be effective in mum fuel exposure obtained was 0.5 at.% burnup reducing the inter-granular attack on tantalum. of the plutonium at a maximum specific rating of Subsequent tests, however, indicated that fuel made 60 W/g. Because of the concentrated nature of the with pure iron was less corrosive than fuel with fuel, these modest figures do not emphasize the sig- added impurities. Likewise, the pure system has the nificant aspects of the fuel experience that had been advantage of reducing the froth region formation obtained. with its resulting reactivity loss. The fuel alloy used in the Mark-I and Mark-I1 cores has a high pluto- The LAMPRE operation successfully demon- nium concentration and is not particularly suitable strated satisfactory fuel performance at levels of up for high performance reactors because of the heat to 1 kW/cm3 of fuel. This figure, translated to removal limitation. Attention has therefore been more dilute fuel (4-g plutonium/cm3) more appro- given to ternary alloys, Le., plutonium-cobalt- priate for larger reactor systems, predicted no prob- cerium and plutonium-copper-cerium, which per- lems from fission gas evolution at 250 W/g. The mit plutonium concentrations from 2 to 4 g/cm2 at fuel exposure level of 2.4 x 1020 fissions per cubic

A-62 15.215.0 05 100 200 300 400 500 600 700 800 Temperature ("c)

Figure A-49. LAMPRE-I plutonium-iron fuel density vs. temperature. centimeter also translates to the same fission prod- designed to have physics characteristics similar to uct density present with a 2.5 to 3 at.% burnup of those calculated for large fast ceramic power reac- the plutonium in the dilute fuel. Thus, with a com- tors. SEFOR was also designed to be operated with paratively modest and inexpensive first experiment fuel composition, temperatures, and cryhtalline with molten plutonium fuel, highly satisfactory states characteristic of proposed fast ceramic power control, reactivity behavior, and successful gas dis- reactors. SEFOR was intended to provide experi- engagement from the fuel was demonstrated. In mental data to promote understanding of the addition, there were no significant effects observed behavior of the Doppler effect under operating on the containment from appreciable concentra- reactor conditions. It was recognized early in the tions of fission products in the fuel. planning of SEFOR that, in addition to the general concern associated with the design of a then rela- SEFOR tively undeveloped reactor system, such as the sodium-cooled fast reactor, unique problems would Concept. On March 27, 1964, the Atomic Energy be involved in SEFOR because of the nature of the Commission, Karlsruhe Laboratory at West experimental program. Requirements for suppres- Germany, EURATOM, Southwest Atomic Energy sion of fuel and core expansion coefficients 1.0 per- Associates, and the General Electric Company for- mit accurate measurements of the Doppler mally embarked on a program to construct and coefficient and forced stability of the core during operate a 20-MW ceramic-fueled, sodium-cooled, the entire testing program, including the power fast flux reactor. The test program to be carried out transient experiments, strongly influenced the with this reactor was intended to verify calculations design of the core and vessel internal hardware. that would indicate that fast ceramic power reactors can be designed to be stable and have desirable Reactor Building. An elevation view of the reactor operating characteristic^.^-^^-^^ The Southwest building is given in Figure A-50. The reactor build- Experimental Fast Oxide Reactor (SEFOR) was ing is divided into two containment barriers. The

A-63 n

Figure A-SO. SEFOR reactor building, elevation view. n

A-64 outer containment structure is an all-welded steel heads. The center head contains a fuel clxe that cylinder 15.2 m in diameter and 32.2 m high, permits several fuel rods to be removed for inspec- designed for a pressure of 2.63 x lo3 Pa. The outer tion without the need for removing the cent’xhead. containment region is normally accessible. Coolant System. A simplified diagram of the The reactor equipment is located below grade in SEFOR coolant system is shown in Figures A-53 a refueling cell that extends above grade. The reac- and A-54. The system comprises two parallel tor core is contained within a steel liner which loops, each consisting of two loops in series. The forms an inner containment barrier in the event numbers in the figures are the elevations of the indi- there is a release of fission products. The area cated points relative to the core center. The primary below grade contains a nitrogen atmosphere during sodium loop transports reactor heat from the reac- operation, whereas the refueling cell contains high tor vessel to the intermediate heat exchangers. The purity argon. The inner containment barrier, was secondary sodium loop, in turn, transfers the heat designed for a pressure of 600 Pa. to the air blast heat exchangers, where it is dumped into the atmosphere. The main coolant loop was During operation of the reactor, the first con- designed for a coolant flow of 315 L/s, whereas the tainment barrier is sealed in the sense that all pene- auxiliary loop was designed for 15.8 L/s. trations to and from this area contain valves or Type 304 stainless steel piping was used throughout doors in the “closed” position. Fluids required, the coolant system. such as nitrogen and argon, are transferred by pairs of batch tanks arranged such that when fluid is The loops were designed so a rupture in one loop being transferred from a batch tank to an area would not affect the operating ability of the other. within the first containment barrier, the batch tank This was done by locating the vessel’s main coolant is valved off from external supplies of fluid. Simi- system nozzles at an elevation such that a break in larly, when the batch tank is being charged with the main loop could not drain the vessel’s !;odium fluid, it is valved off from the inner containment below a point well above the reactor core. To guard area. A similar scheme is used to remove waste gas against leaks in the reactor vessel itself, a safkty ves- from the inner containment. The nitrogen within sel in the form of a cuff was placed around the the first containment barrier is cooled by means of vessel in near the core as shown in Figure A-54.The freon refrigeration units arranged so that only safety vessel extends up to the main coolant system closed coils penetrate the first containment barrier. nozzles, keeping the sodium level well below the The air region between the steel-lined cells and the auxiliary loop dip tubes in case a vessel leak outer containment vessel is also cooled by a freon develops below this elevation. The safety vessel is refrigeration unit, but it is normally ventilated via a welded and hermetically sealed to the reactor ves- 15-cm air intake line, and a 15-cm air exhaust line sel. Provisions were made to periodically pull a vac- that exhausts to the atmosphere through a small uum on the volume between the two vessels to stack. verify the leak tightness of the safety vessel. Sodium leak detectors continuously monil or for Reactor ve.sse/. The general arrangement of the ves- sodium leakage into the safety vessel. sel and internal structure is shown in Figures A-51 and A-52. The vessel consists of a lower, small Precautions were taken to minimize probable diameter core section and an upper, large diameter loss of pumping power during an accident. Because shielding section. Design pressure and temperature electromagnetic pumping systems have low inertia are 5.3 kPa and 570°C, respectively. The reactor characteristics, the sodium coolant flow would will normally operate at a positive pressure of drop suddenly if pumping power were lost. To pre- approximately 15.2 cm of water pressure relative to vent this, two independent motor-generator sets the refueling cell. The vessel head is designed in two with built-in flywheel inertia units were used to pieces: the outer peripheral head which is seal- power the main primary coolant system pump. welded to the vessel flange and the center head Each set is connected to one of the two coils operat- which is normally held in place by its static weight. ing the pump and is driven by the external power The center head can be removed for refueling the obtained from the incoming 69-kV line. If external core. The outer head needs to be removed only in power is suddenly lost, these sets coast down on the event that major repairs are required in the core their flywheel inertia and maintain sufficienlt flow support structure. Special accident hold-on struc- to prevent an increase in the coolant temperature. tures are provided for both the outer and center The flywheels were sized such that the coolant flow

A-65 CR1 N1 thru thru CR6 91 112 N6 27 ?3 dia.4 \-p I I

1631 ft-10 in.

1630 ft-2 in. - 48 dia.' --

_A_C.-.

4ux. hield 1627 ft-9 in.

i 1 I! 7. of 1625 ft-10 in. innels

r 1 80 114 1.-

4 1.- ef!e:Cto

35 core - --

.I .5 - 1 .. - ,eflecto - --1 .. - ~-7-- ~------. --I==

E42 112 dia.---

Figure A-51. SEFOR reactor vessel, elevation.

A-66 Figure A-52. SEFOR reactor vessel, plan view.

ux pump 5

Approx. elevations in feet with respect -4 to core - Piping lengths (feet) Main Aux Reactor to IHX 50 34 IHX to pump 57 1 Pump to reactor 92 28 t 1 26 Main pump

Figure A-53. SEFOR coolant system.

A-67 Aux. inlet and -.=..outlet are at 1500 and 2100 angular position Main

elevations in inches

Figure A-54. SEFOR reactor vessel nozzle elevation.

is still greater than 50% of full flow 15 seconds tem ceased, and if all power were lost to the plant at after loss of power. the same time, the auxiliary system’s natural circulation could remove the decay heat to stop the Before a significant rise in coolant temperature sodium in the core from boiling. can occur, simultaneous loss of the two motor- generator sets or the two independent coils of the Refueling Cell Facility. The refueling cell facility EM pump has to take place. Even if such simulta- consists of a concrete and steel structure within the neous failures occur, calculations indicate that the reactor building. The cell provides remote access to coolant temperature will not exceed 704°C when the fuel elements and other reactor internals. In the reactor is scrammed from full power addition, facilities were included for fuel storage [20 MW(t)]. inspection and preparation for shipping.

During periods when the reactor vessel or fuel To further increase reliability of the coolant sys- storage tank is open for access, or when fuel or tem, the main coolant system, as well as intermedi- other irradiated core components are exposed in ate and air blast heat exchangers, was located such the cell, the cell serves as the first containment bar- that decay heating can be removed by natural con- rier. During normal operation, this cell controls the vection. Thus, shutdown cooling is possible even if transfer of gases, vapors, and particulate matter no power is available. between the cell’s interior and all other areas.

The intermediate and air blast heat exchangers of Access by personnel to the cell was minimized the auxiliary system were also elevated to provide and only possible when the reactor was inoperable n natural circulation. If flow in the main coolant sys- with all irradiated fuel adequately shielded. A-68 When no alpha contamination was detected The reflector controls are calculated to be worth within the cell, its argon atmosphere was main- $14, thus leaving a margin in excess of $5. The tained at a positive 2.54 f 1.25 cm of water rela- reflector worth was measured in a SEFOR critical tive to the operating room. This was to minimize experiment at ZPR-111. in-leakage of oxygen. During periods of alpha contamination, the cell atmosphere was maintained in Reactor shim control was accomplished by mov- a -5 + 2.5 cm of water relative to the operating ing the reflector control rods vertically at a constant room to prevent out-leakage of contaminants. velocity into positions adjacent to the core. The reactor was scrammed by simultaneously dropping A 10-ton multi-speed ridge crane installed in the all the reflector rods to a position below the core. cell was used to perform most of the handling oper- The general arrangement of the reflector control ations, including the handling of the fuel rods. Spe- system is shown in Figure A-55.The reflector was cial grapple tools were provided for handling 15 cm thick and 86.4 cm high with the annular cyl- various components. Manipulators were used to inder divided into 10 equal sectors. The average assist in the handling motions. worth of each sector is $1.4.

The cell had an argon atmosphere and was The control rod assembly consists of a reflector designed to withstand 2.62 x lo2 kPa. The argon structure in a neutron shield. The control is was was circulated through a purification system to moved within the reflector guide, which extends maintain less than 10-ppm oxygen, 10-ppm mois- from the ceiling of the drive mechanism cell to the ture, and acceptable sodium vapor conditions. The top of the core. The channels in the reflector guide purification system consisted of a NaK bubbler and both guide the rods and direct the flow of coolant demister which controlled the impurity limits in a gas over the rods. Each reflector rod is connected to NaK exchanger and maintained the temperature its drive by an extension shaft. below the 66°C design temperature. The reflector material chosen for the reference Reactivity Control. SEFOR was controlled by design is “A” nickel, a high nickel alloy containing movable reflector control rods surrounding the approximately 99% Ni. This alloy, though not a core outside the reactor vessel. The reflector con- high strength material, has a relatively good ther- trol provides a cleaner core and is more reliable if mal conductivity, which is effective in milnimizing structural damage should occur to the core during thermal stresses and distortions. The shielding is the transients. Control requirements for the contained within a steel can with the same geome- SEFOR are summarized in Table A-7. try as the reflector. Slider pads fit into channels on

Table A-7. Reactivity control requirements

Reactivity ($1

Temperature (177OC to 29 MW)

Doppler 2.7 Expansion 1.6

Experiments 2.0

Burnup 0.3

Shutdown 2.0

8.6

A-69 -143 in. floor level

T Core 34 in. Reflector

Vessel 22 1124 01 Instrument tubes (1 2)

Plug; Reflector shaft

-- Coupling .- Drive control area - Twin screw drive - Coarse drive T-1 ft-5 in. --.Cylinder (40 in. stroke)

--- Position indicator 17 in.1 -- Motor dirve

n123456 Iy Graphic scale in feet .12ft-0 in.

Figure A-55. SEFOR reflector control system.

A-70 the faces of the control rods. The control rods are move the stop down to maintain contact with the on rails within the reflector guide. The estimated cylinder rod. A speed control orifice in the hydrau- weight of each control rod is 660 kg. lic supply line limits the rising speed of the rod in the event of an accidental pressure reversal after Two types of drive mechanisms are used, a fine disengagement from the stop. control and a coarse control. The fine control drives are infinitely positionable over their full The coarse control drive consists of the hydraulic stroke, whereas the coarse control drives are nor- cylinder alone, because the twin screw positioned mally operated fully inserted or fully withdrawn. A stop is not required for this drive. Rod position is twin screw drive, illustrated in Figure A-56, is used indicated by a selsyn transmitter geared to a rack as the fine control drive, and has a record of satis- and attached to the cylinder rod. The cylinders for factory performance. both the fine and coarse control drive are identical. A shock absorber is mounted in the lower c,ylinder The fine control drive consists of an oil hydraulic head to decelerate the control rod following a cylinder connected directly to the control rod by an scram. extension shaft. The cylinder pressure is normally biased so that the cylinder tends to move the rod in core Design. The SEFOR core structure is made the direction to increase reactivity. However, the up of 109 hexagonal channels resting on a support cylinder motion is restrained by an adjustable stop, plate. Each channel is fastened to the support plate which consists of two translating nuts driven by two by an orifice that is lowered in the channel, inserted synchronized screws. The stop is positioned by a through the support plate, and rotated 90 degrees three-phase gear motor through a drive chain that to lock. Fuel and tightener components resting on drives and synchronizes the screw. Stop position is the orifice prevent unlocking during operation. indicated by the selsyn transmitter geared to the One-hundred nine channels are clamped around drive train. Scram was accomplished by deenergiz- the periphery of the core, making a circle approxi- ing a scram valve that dumps the supporting oil mately 86 cm in diameter. Each channel contains a pressure to a low pressure reserve. The resulting center tightener rod and six fuel rods, one in each unbalanced force causes the cylinder to retract from corner of the hexagonal channel. The tightener rod the stop at an initial acceleration of about 1.5 g. is inserted into position after the fuel rods are in The rod is decelerated by a hydraulic buffer incor- place. Fuel rods are individually removed and porated in the cylinder head. A limit switch in the replaced. Channels are not removed from the reac- adjustable stop automatically signals the drive to tor during normal servicing. The channds are

Ref'ector sha\ft Limit switch Cylinder shaft ,- Limit switch Cylinder \ 7

Section AA I-7 in.-10 in. (approx.) Coarse control drive

Ref lector shaft Clyinder shaft - Twin drive Motor drive shaft Position indicator , Limit switch Limit switch / [assembly Cylinder 1 /~r\ven iotor '> mit switch Section AA (scra 4 L-9 in.-2 in. (approx.) 4 Fine control drive

Figure A-56. SEFOR reflector control drive.

A-7 1 approximately 254 cm long. The top of the channel The mock-up represented four main sections: is 20.3 cm below the top of the fuel rods to provide for the routing of instrument leads from the instru- e Core

mented fuel rods. .I Radial and axial reflector The fuel rod assembly consists of two pieces: a fuel rod and an extension rod. The two sections are Sodium and steel structure between core pinned together to form the fuel rod assembly. and reflector

The active fuel column in the core is 85.9 cm Boron carbide neutron shielding sur- long. A 10.2-cm nickel reflector is located above rounding the radial reflector. and below the fuel. Both fuel and reflector are contained inside Spe-316 stainless steel tubing, 2.54-cm Experimental Program. The experiments per- OD, 1.4-m wall thickness. The lower end of the fuel formed at SEFOR were to obtain power coefficient rod fits into a rod sheet in the channel assembly. information applicable to large fast power reactors The upper end provides a connector for attachment at steady state conditions and during fast tran- to the extension rod. sients. Experiments between these two domains, i.e., for slowly varying oscillatory operations, were The fuel extension rod consists of borated graph- also carried out with SEFOR to provide further ite pellets contained in 2.22-cm-OD tubing. A information. The goal of the SEFOR test program socket at the upper end provides a grapple tool was to make it possible to predict with confidence attachment. The upper end of the extension rod the safety characteristics of future large fast and fuel rod are identical so a single grapple tool ceramic power reactors. can handle either section after detachment. Splined regions enlarged the diameter to 2.54 cm at eleva- It was shown that a low-power experimental reactions where the tightening springs are needed. tor, with large-diameter fuel rods, can give informa- Splines are necessary for passage of adequate tion on power coefficients directly applicable to coolant. high-power reactors with small-diameter fuel rods under steady state and fast transient conditions .A-58 Development Testing. Early in the planning of the SEFOR program it became obvious that several SEFOR experiments indicate that the reactivity areas of the reactor would need development test- effect caused by the change between two steady ing to ensure satisfactory performance. The variety state power levels with unchanged fuel clad temper- of power transients, ranging from oscillatory exper- ature is independent of fuel rod radius. This reac- iments to prompt power excursions, placed unusual tivity effect is the product of the change in power operating conditions on the reactor components. It per unit length of rod and a parameter, G, or was desirable to have information on core behavior to help interpret the experiments during operation. Ak = GA(P/nH), (A-7) Consequently, a development test program was planned covering critical experiments: in-pile and where n is the number of rods, P is the power, and out-of-pile testing of fuel assemblies and core; H is the rod length. For a given fuel material, refueling cell mock-up; instrumentation; and spe- G depends only on P/nH and the fuel temperature cial reactivity controls. coefficients.

A series of critical experiments was conducted A similar result was obtained at SEFOR for fast with a mock-up of the SEFOR core and the ZPR- transients during the period when power is chang- 111 critical facility at the Idaho National Engineer- ing fast enough that heat conduction of the fuel rod ing Laboratory. Measurements included the can be ignored. For this case the time-dependent plutonium requirement, Doppler coefficient, and feedback reactivity is reflector control strength. Results of these experiments were used to check the calculational techniques used in the SEFOR analy~is.*-~~-~~ Ak = YA(E/Vf),

A-72 . ..

where E is the excess energy produced during the structure is included in the coolant cocfficient excursion to time t and Vf is the fuel volume. For a because the structure temperature follow the given fuel material, Y is a function only of the fuel sodium temperature. Radial expansion is not con- temperature coefficient (Po/nH), and (E/Vf), trolled by the grid plate in SEFOR so a grid plate where Po is the power level at the start of the excur- coefficient was not included. sion. Again, the radius of the fuel rod did not enter explicitly. The combined fuel and coolant uniform temperature coefficient was measured at zero power by Three types of steady state tests were made at externally heating the sodium. Provisioiis were SEFOR. The purpose of these was to measure the made to heat the sodium to 540°C. The response to following: reactivity oscillations from zero power to high power was measured to determine the feedback The power coefficient associated with transfer function. At low frequencies, the feedback changes in fuel temperature only reactivity provides a measure of a steady state Parameter, G [Equation (A-7)] which involves the The coolant temperature coefficient Doppler coefficient and the fuel conductivity, k. These oscillator results in effect duplicate and ver- The combined fuel and coolant tempera- ify the steady state measurement. At high frequen- ture coefficient cies the feedback reactivity provides a measure of a transient parameter [Equation (A-8)], which The flow coefficient and effect of minor involves the Doppler coefficient and the heat changes in structure geometry. capacity, or pCp. The high frequency results provided an independent measurement of the parame- The first test was the reactivity measurement ters important to the transient behavior of the described by Equation (A-7). The reactivity effect reactor and permitted a prediction of its transient caused by changing the steady state power level performance prior to the transient tests. At inter- while holding the average coolant and structure mediate frequencies, the feedback reactivity is temperature constant was measured. The average characteristic of neither steady state nor fa:a tran- coolant temperature was held constant by variation sient operation alone; the results depend on both of the primary flow rate and variation in secondary k and pCp. coolant conditions. The channel and Be0 temperatures follow the coolant temperature. Second order Fast transient tests were made to obtain informa- corrections were made for small changes in average tion about the energy coefficient Y of Equa- clad temperature. In these steady state tests, the tion (A-8), a parameter relating fuel tempera.ture to Parameter, G,of Equation (A-7) was measured as a the heat capacity. The transient experiments came function of P/nH for the SEFOR oxide fuel up to after a thorough analysis of the steady state and P/nH equals 11 kW/ft, the value at 20 MW. oscillator tests.

To the extent that the thermal conductivity of the Both subprompt critical transient (Ak < $1) SEFOR fuel was known, the Parameter G could be and prompt critical transients were run. Reactivity related to the fuel temperature coefficient. SEFOR ramps were introduced by injecting poison from the fuel was designed to minimize the axial fuel expan- center of the core by means of a fast reactivity sion coefficient so that the Doppler effect accounts excursion device (FRED). Maximum ramp rates for most of the fuel temperature coefficient. In this and maximum values for FRED depend on the val- way, the first type of steady state test provides a ues obtained for the Doppler coefficient from the measurement of a Doppler coefficient. steady state and oscillator tests. The transient test ,1 program was not intended to approach the regime The coolant temperature coefficient can then be where fuel damage would result. measured independently of fuel temperature effects. While operating at a fixed power level, the Demonstration of Equipment Reliability in Operating reactivity effect caused by changing the average Lifetime. Two designs measured the sodium kvel in sodium temperature is measured. The sodium tem- the SEFOR reactor vessel. The reliability of either perature can be changed either by altering the pri- design was satisfactory for short periods, 5:everal mary flow rate or by altering the secondary coolant months, the number of months depending on serv- conditions. The effect of radial expansion of the ice conditions during the time. However, without

A-7 3 frequent cleaning or special care, the performance reactivity, (c) measurement of temperature and of either was inadequate for operating periods of power coefficients, (d) determination of spatial several years .A-59 neutron-energy distribution and heat generation, (e) measurements of neutron flux in experimental During the more than three years of operation, channel sections of the reactor, and (f) determina- the electromagnetic sodium pumps operated with- tion of radioactivity effects and safety aspects. out a malfunction. Some malfunctions of pump power supply and controllers did occur, but these The general arrangement of the reactor is shown components were easily and quickly repaired. The in Figure A-57. An unusual feature is the horizon- reliability of fuel grapples for nearly 4,000 fuel tal graphite column, 3.1 m long, which penetrates transfers was very good. one side of the shield and provides a number of locations for horizontal and vertical experiments. USSR BR-5 Maximum power density was 500 kW(t)/L corresponding to a maximum heat flux of 1.39 MW/ The first USSR fast reactor, BR-1, was built in m2. 1955 as a zero energy assembly; it was fueled with plutonium and was used to investigate fast reactor Core Assembly. The core consisted of 88 hexago- physics. The work was extended by BR-2, a pluto- nal fuel assemblies, 2.6 cm across flats and nium filled, mercury-cooled reactor built in 1956 83.8 cm long, surrounded by two rows of blanket and operated to 100 kW(t). The BR-2 provided the subassemblies. Each fuel subassembly, shown in facilities for physics experiments and irradiation of Figure A-58, contained 19 stainless steel clad pins materials in fast neutron fluxes at 1014 neutrons/ filled with plutonium oxide pellets and sealed after cm2/s, and furnished experience in operating at they had been charged with helium gas. The pins temperatures to 160"C.A-60361 were 0.5 cm OD. The fuel itself was about 0.41 cm in diameter, and the cladding was 0.041 cm thick. The BR-2 was dismantled in parts, such as shield- The pins were separated by spiral wire spacers. Sub- ing, for use in construction of the BR-5 reactor assemblies were supported at the bottom by a per- which was completed in 1958. It went cold critical in forated plate. The core subassemblies and a ring of the summer of 1958 and was in operation the sum- tubes loaded with natural uranium were enclosed in mer of 1959. The BR-5 operated at a power of a stainless steel vessel approximately 36 cm in 5 MW(t), was fueled by plutonium oxide, and had diameter. Clearance between subassemblies was sodium inlet and outlet temperatures of 375 and 0.05 cm. The core itself was approximately 28 cm 450"C, respectively. The main purpose of the reac- by 28 cm. There was no top plate, and no positive tor was to gain burnup data on fuel elements, to lateral restraint was provided at the center or top. gain experience operating radioactive sodium cir- Sodium flowed up through the core from a single cuits and to irradiate various materials. Since it was inlet at the bottom of the central sodium tube to a not considered necessary to obtain further data on single outlet above the core that supplied two cool- breeding gain, the design was simplified by omit- ing circuits. ting the normal blanket of natural or depleted uranium .A42963 Control Elements. Movement of an inner and outer cylindrical reflector was used to control the The reactor started up at zero power July 1958. reactor. The inner nickel cylinder, 5.1 cm thick, was From July 1958 to January 1959, the assembly of raised or lowered by cables; the outer 10-cm-thick the primary loop was completed and startup modi- nickel cylinder had two movable sections that pro- fications were accomplished. In January 1959, the vided fine control. Emergency shutdown was reactor was started up with sodium coolant. From accomplished by electromagnetically releasing January to July 1959, the reactor operated to these two cylinders. In addition, two controls oper- 1 MW. On July 21, 1959, full power of 5 MW(t) ating in the outer cylinder provided shim adjust- was reached. Later, the reactor was operated at var- ment. Heat generated in the reflector was removed ious levels to 5 MW(t) to obtain a 450°C sodium by air. Startup instrumentation was located in the outlet temperature. The sodium temperature in the water tank outside the reflector, as shown in primary loop was raised to 500°C in December Figure A-57. 1960. During 1960 and 1961, the reactor was used for experiments: (a) calibration of control rods, When the assembly was heated from 40 to (b) determination of circulating sodium's effect on 600"C, the radioactive gas release rate increased by

A-I4 Figure A-57. General arrangement of the USSR fast reactor BR-5. a factor of lo3 to lo4 times. During the course of elements, having a diameter of 76.5 cm and a testing in a storage hood, the two assemblies previ- height of 38.0 cm. The fuel in the element was ously having activities below background level encased in stainless steel cladding. Clearance space showed very marked gas release when they were between the fuel and cladding was filled with heated to 500°C. Heating to 700°C revealed a high helium, and stainless steel wire 0.4 m thick was activity in the majority of assemblies. To be certain used as a spacer. A space 50 mm high was left above no irreversible changes occurred in the fuel clad- the fuel elements to collect gaseous fission prod- ding as a result of heating to 700"C, two of the ucts. The calculated maximum temperature of the seven assemblies showing marked activity after fuel at the center of the elements was 800°C. heating to that temperature were retested at 40 and 400°C after cooling, with no major gas release Startup of the monocarbide-fueled BR-5 reactor being observed. took place in May 1965. The fuel element :assemblies were loaded in the active core zone of the reac- The fuel used for the second loading of the BR-5 tor which had previously been filled with sodium. reactor's active core zone was uranium monocar- Criticality was reached when 83 assemblies of ura- bide with a 90% U-235 enrichment and an over- nium monocarbide fuel elements (99 kg IU-235) stoichiometric carbon content. The density of the had been loaded together with 20 blanket assem- uranium monocarbide was 12 g/cm3. The original blies containing natural uranium metal of the origi- dimensions of the fuel element assemblies were nal design. The critical mass obtained agreed fairly retained, but the design was changed. Each assem- closely with the data acquired using tht: two- bly in the new active core zone contained seven fuel dimensional calculation in the 18-group diffusion

A-75 approximation, which predicted a critical mass of 89 assemblies.

To conduct further studies on fuel burnup in the reactor, four assemblies of plutonium oxide elements were loaded; these elements previously had been in the facility and had reached a burnup of approximately 6%. At the same time, two assemblies of plutonium dioxide elements and two assemblies of uranium dioxide elements with a burnup to 1% were inserted in the reactor. The plutonium dioxide was also used as a startup neutron source by using the (alpha, n) oxygen reaction and the spontaneous fission of plutonium-240.

Section A-A After the fuel had been loaded, studies were carried out on the reactor characteristics. Heat generation over the radius and height of the active core zone was calculated on the basis of gamma activity distribution in the fuel element assemblies that were mounted at different radii. The effect of the drainage of sodium from the active zone was measured by emptying all the sodium from the central reactor pipe under subcritical conditions. The reactor temperature coefficient, which is related to the uniformed heating of the elements in the active zone and the increase in the temperature of the sodium at the reactor inlet, was measured at different power levels ranging from 10 to 4000 kW. The temperature coefficient was found to be constant for sodium inlet temperatures ranging from 180 to 430°C. The steady state power coefficient, which is a function of the nonhomogeneous heating of the reactor elements, was measured on the basis of the change in reactivity as power increased at the constant sodium inlet temperature. The steady state power coefficient was constant throughout the power range measured. The reactor power effect was measured by observing the reactivity curve after power was stepped up by 10% at a starting value for various sodium flow rates. In each measurement, the inlet temperature of the sodium rose by 75°C.

Two temperature coefficients of reactivity were measured in BR-5, one relative to the inlet temperature and the second relative to the outlet temperature. These agreed with predictions and were both negative. The fuel elements, supported .in a lower support plate, were free to move at the top to take up any tolerances in the lower plate. The pins within the subassemblies were welded together at Fuel pin the bottom. Alternate pins were wound spirally with stainless steel wires, which acted as spacers. Figure A-58. BR-5 fuel assembly. Bowing of pins in the subassemblies tended to

A-76 .. .. - - .- .. -.. . -. .. . - . . .- ......

cause the outer element tubes to move outward to normal elements. A 2.54-cm-diameter reentrant reduce reactivity. An increase in coolant flow had a tube in the center of the core, designed for testing similar effect. fuel elements, was cooled with a separate NaK cooling system. In this central loop, a peak flux of Coolant System. The coolant system was split into neutrons/cm2.s was available. A flux of one- two IO-cm-diameter primary loops. The pump seal half this value was available in the exposure holes at and bearings operated in argon cover gas continu- the end of the core. ously circulated through a silica gel dryer and then through copper-coated silica gel heated to 250°C. Operating Experience. Pump life was mainly deter- The totally enclosed pump motor was likewise mined by ball bearing capabilities; it averaged maintained under an argon gas pressure slightly 8,000 h. Replacement of primary equipment, higher than that of the system, so any possible leak- including cold traps, usually started a week after age was through the gland into the system. shutdown following sodium gamma dlxay. The cold traps in the primary system were repllaced four Heat was transferred from the primary loops by times during the first 4 years of operation. two intermediate heat exchangers to two secondary loops. One of these NaK-filled secondary loops In October 1960, with about 2.5% inaximum was provided with a 1.69-MPa steam generator plutonium burnup, the leakage of fission products having duplex tubes. Mercury was used as a heat into the sodium became pronounced. By transfer medium in the space between the inner and September 1961, the cesium activity in the sodium outer tubes. An air-cooled heat exchanger and a was 70% of total cesium. At this time, the fuel blower, delivering 11 m3/s of air, were used to burnup had reached 4.85%. The reactor was reject the heat in the other loop through a 24.4-m unloaded, and the subassemblies weie steam- stack. Residual heat after shutdown was also blasted clean. During cleaning, some of the subas- removed in this loop by natural convection. semblies showed considerable increase of radioactivity in the steam condensate owing to Shielding, Components, and Structures. A water increased failure of fuel pin cladding. Steam clean- tank surrounded the core and provided a 51-cm ing was discontinued, and a leak test, with an ion- shielding annulus. Surrounding this was a 41-cm- ization chamber, was then made to check each thick cast iron ring followed by a concrete wall with subassembly to measure fission gas activity. Of the an average thickness of 109 cm, as shown in 81 subassemblies tested, 8 had been washed with Figure A-57. Composing the upper shield were a steam; of the remainder, 63 showed normal sodium boron carbide plug, two rotating steel plugs, and impurity activity, and 10 had a gas activity higher finally a removable top plug (consisting of layers of by a factor of 1,000. Examination of the 4.85% parafin wax and iron). burnup pins showed longitudinal cracks in the cladding caused by fuel swelling. The cracks showed up The lower shield consisted of a 20-cm-deep water with a clearance between cladding and fuel that was layer, which allowed entry to the cell under the reac- at a minimum. tor only during shutdown. After fuel removal, the primary system was Fuel loading was accomplished through a double drained clean with steam at 'LO. 1 MPa and %120"C eccentric rotating plug arrangement. During refuel- by injecting the steam into vent lines and removing ing, the reactor was shut down, and sodium circula- the sodium from the low points in the reactor sys- tion in a primary loop was stopped. Decay heat was tem. The system was kept at about 150°C' prior to removed by natural convection. A transfer cask unit steam injection. About 9,000 kg of steam was was used for charging fuel, uranium, and experimen- used, and no violent reactions occurred. Contami- tal units. The reactor had no outer containment build- nation of the primary system was reduced 50%. ing. The primary loops were housed in separate The system was then filled with pure water twice, concrete cells, and the secondary circuit equipment with no circulation, and drained. There was no was contained within a single compartment. decrease in activity. Each section of the primary system, excepting a nickel basket in the real:tor, was Experimental Facilities. The reactor had been then cleaned with a solution of 5% nitric: acid at designed not only to permit sample radiation under about 70"C, with three successive flushes to reduce fast flux conditions but also to allow the insertion the activity by a factor of two to three. N[oreover, of special fuel subasse\mblies in the core in place of two sections were subjected to a 0.5% KMn04

A-77 solution flush for 24 h at about 70"C, a 5% nitric increased considerably. The radioactive contamina- acid and 1% oxalic acid mixture flush for 3 to 4 h tion of the primary circuit, after the decay of the at 70"C, and a pure water flush for 1 h at about sodium-24, consisted of relatively short-lived fis- 70°C. The three flushes were repeated five times. sion products: iodine-13 1, barium-140, Activity was reduced to a value low enough so that lanthanum-140, cesium-136, and cesium-137. The repair work could begin. A total of 21 flushes per activity of the iodine-131 was greater than that of section were performed, including the steam flush. the other isotopes amounting to 30 mCi/g of This indicates the possible consequences of plating sodium. At the same time, it was found that the out fission products on system components. Radio- amount of iodine-131 released from the leaking chemical analysis showed plutonium, zirconium, fuel elements was usually lower than that of the and cesium as the residual activity. After the final cesium-137 release. Fission products such as drain, the primary system was dried using vacuum iodine- 13 1, cesium- 136, and cesium- 137 were effi- and heat. Following repairs the system was filled ciently retained by the oxide cold trap. For exam- with distilled sodium. The reactor was placed back ple, a sodium oxide cold trap that had been in use in operation in March 1962 after a 7 month shut- for 2 years, was cut out of the circuit in the middle down. Eighty percent of the original bundles were of 1964. The total activity of the cesium-137 in this placed back in the reactor. Subsequent reports indi- cold trap was found to be 10 times that of the cate some of these began to leak soon after cesium then present in the primary ~ircuit.~-~~ startup.A-62-65 The cold trap did not possess any great efficiency After March 1962, the charge in the BR-5 active with regard to the retention of isotopes such as core zone was replaced by plutonium dioxide and zirconium-95, niobium-95, barium- 140, and uranium dioxide assemblies. For further operation, lanthanum- 140; these radionuclides were mainly the reactor was loaded with 80% plutonium dioxide deposited on the walls of the equipment and the assemblies from the original charge, with a maxi- piping. In 1964, when the reactor was running at mum burnup of 4.85% and uranium dioxide 3.5 MW, special studies were performed to examine assemblies of similar design. Pu02 assemblies were the influence of the cold trap on activity of the fis- concentrated in the central part of the active core sion products in the coolant and on the walls of the zone. While loading the assemblies they were rear- primary circuit piping. With the aid of of a gamma- ranged in order to equalize their burnup during ray scintillation spectrometer, measurements were subsequent operation of the reactor.A-66 made on the bypass section of the primary circuit. The measurements showed that when the cold trap During 1962 and 1963, BR-5 operated at 20 and was working, activity of the coolant in the pipe 40% of its rated power. In 1964, the output was wells attributable to cesium-137 and iodine-13 1 stepped up to 3 MW and then 3.5 MW, set equal to were only one-tenth of that when the cold trap was 70% of the rated power. After operating 3 years not in use. with fuel elements that had reached high burnup and had leaks in the cladding, it was possible to Activity of gaseous fission products in the gas acquire some experience and information about the spaces markedly depended on the power level. At a contamination of the coolant, inert gas, and pri- power of 1 MW, the xenon-133 and xenon-135 activ- mary circuit equipment, because of the radioactive ity was of the order of a few tenths of a millicurie per fission products. It was found that the release of liter; however, at 3 MW it amounted to 20 and 5 mil- gaseous and volatile solid fission products from the licuries per liter and at 3.5 MW, to 250 millicurie per fuel elements to the coolant was mainly dependent liter, respectively. Despite different conditions for the on the reactor power, and therefore on the tempera- release of gas from the sodium to gas spaces in the ture of the fuel elements. At a power of 1 MW, circuit (pumps, central reactor pipe) the average there was virtually no release of fission products xenon-133 activity in the gas blankets was roughly the from the fuel elements, and a contamination of the same. Yet the ratio of xenon-133 to xenon-135 in them circuit was found to consist of Cs-137. The Cs-137 was quite different. in the coolant was approximately 0.3 mCi/g of sodium and was caused by the residual contamina- At a fairly high iodine-13 1 content in the coolant tion of the primary coolant; in addition, the quan- and on the walls of the flow equipment, the release tity of cesium removed from the sodium by the cold of iodine-131 to the gas spaces in the primary cir- trap was extremely small. At a reactor power level cuit was not very marked. The iodine-131 activity of 3.5 MW, the release of fission products in the gas did not exceed millicuries/liter,

A-78 which was lower than the xenon-'133 and xenon-135 tion was localized in the event that any fuel ele- activity by a factor of lo3 to lo4. ments in the assembly were defective. The assemblies removed and placed in molten lead con- Measurement of circulating gas activity revealed sisted of 8 plutonium dioxide assemblies with a the presence of neon-23 [formed from sodium-23 maximum burnup of 6.2% and 2 uranium dioxide by the (n,p) reaction] in the inert gas of the primary assemblies with a maximum burnup of 1.4%. A circuit. Total activity of the neon in the circuit's gas month after fuel unloading, a check was conducted spaces was about 50 curies. Circulating sodium on the leak tightness of the fuel elements' cladding activity in the bypass section was measured a few and all the assemblies in storage. Gas was drawn off minutes after reactor power was increased because into an ionization chamber through a pipe at that time there would be low sodium 24 activity. mounted in a space between the hood of the storage It was therefore possible to determine the neon area and the bundle in it. The temperature of the activity in the coolant which was equal to 0.6Ci/L assemblies was determined on the basis of the after- or about 1,000 curies for the entire circuit. heat (approximately 40 watts per assembly) and the temperature of the surrounding air. The activity of In November 1964, additional leaks in the pluto- the gas drawn off the assemblies provided an indi- nium dioxide fuel elements resulted in major cation of the leak tightness of the fuel cladding. In increases in gaseous and solid fission products in this way, checks were carried out on 59 assemblies the sodium circuit. Consequently, BR-5 was shut containing plutonium dioxide element,;. with a down to allow loading of monocarbide fuel into the burnup of up to 6.5'70, and 21 assemblies contain- active core zone. At the time the reactor was shut ing uranium dioxide and uranium monocarbide down for reloading, the maximum burnup in the with a burnup of up to 1.5%. In the COUI-seof the plutonium dioxide fuel element forming part of the check performed, all the uranium assemblies and original charge was 6.5% and that of the 90% 32 of those containing plutonium were found to be enriched uranium oxide elements 1.4%. At this leak tight; activity of the gas released from them time, the fuel elements and some of the uranium did not exceed that of the gas from tht: dummy monocarbide assemblies placed in the reactor in assemblies containing metallurgical sl: ecimens. 1964 to study their efficiency had reached a total These specimens were removed from the active core burnup of 0.8%. Loading the reactor's active core zone and, like the hot assemblies, had not been zone was followed by completion of maintenance washed free of sodium. Gas activity from the work, mainly directed at replacing leaky drainage 17 plutonium assemblies with a burnup of more valves in the primary circuit loops. Cables in vari- than 5.1% was greater than the background by a ous control systems associated with the main cir- factor of 10-1,000. Measurement of a gamma spec- cuits were replaced, owing to their unsatisfactory trum of a gas from the assemblies showed the activ- electrical insulating properties and the need to ity was made up of krypton435 (measurements were attach additional experimental devices to the reactor. performed 4 months after shutdown of the reactor).

In January 1965, the active core zone was entirely To study the active gas released from leaking fuel unloaded in 9 d and placed in the spent fuel storage elements, tests were made on two of the leaking area. Unloading was carried out without cooling of assemblies and 15 assemblies emitting bazkground the assemblies. The sodium and its oxides were not activity. The tests were performed in a special hood washed from the bundles removed from the active fitted with an electric heating element and a ther- core zone. mocouple. In both leaking assemblies, sa marked correlation was observed between the gas release In an attempt to examine the possibility of and the assembly temperature. unloading high activity fuel assemblies into molten lead, 10 assemblies which had just been removed from the active zone were immersed in casks con- Early Power and Test Reactors. The early taining molten lead at approximately 500°C. No power and test reactors were develop1:d in the external signs of a reaction were observed when the Federal Republic of Germany, France, Italy, Japan, sodium on the surface of the assembly dissolved in United Kingdom, United States, and the Union of the lead. The residual sodium was washed off the Soviet Socialist Republic. They were primarily assemblies when they were replaced in the lead; in intended to demonstrate an integrated LhlFBR sys- addition, the source of any radioactive contamina- tem, including the capability for electrical genera-

A-79 tion. Experiments were conducted on the dynamic The core of the KNK-I1 experimental breeder behavior of the various reactor systems under oper- power plant was designed as a two-zone core. The n ating conditions comparable to those existing in a reactor power was 58 MW(t) and 28 MW(e) with a large-scale power plant. A great number of data rod power of 450 W/cm. This first core of KNK-I1 have been gathered with respect to fuel-cladding- has fuel rods with 6-mm-diameter and spark- coolant capability, performance limits of ceramic, eroded grid type spacers. Surrounding the test zone alloy, and metallic fuels, reprocessing and recycling are 22 fuel elements of the dragger zone, as well as technology, basic nuclear data, Doppler coeffi- the five units of the primary and three units of the cients, sodium buoyancy effects, component devel- secondary shutdown systems. opment, instrumentation and control, and sodium technology. These reactors have provided the per- The first core of KNK-I1 was decommissioned formance data needed for the safe design, siting, on August 30, 1982, after an inpile time of construction, and operation of prototype and dem- 400 equivalent full-power days. The fuel elements onstration LMFBRs. The characteristics of the of the test ,zone had-obtained peak burnups up to early power and test reactors are summarized in 100,000 MWd/t, exceeding the design level by Table A-8: nearly 20%.

Federal Republic of Germany Failed Fuel Rod Detection-In KNK-11, failed fuel elements are detected by continuous global moni- KNK-2 Operating Experience. Konpakte Natrigume- toring of the reactor cover gas, of the argon cover kuhlte Kernreaktoranlage (KNK) at Karlsruhe was the gas for fission products, and of the primary cool- first liquid sodium-cooled power station in Germany. ant for delayed neutrons. Gas monitoring is per- KNK was a turnkey project with Gesellschaft Fur formed using a gas chromotograph, a xenon Kernforschung being the owner, Interatom the manu- absorption bed with an activated carbon filter, facturer, and KBG the operating company. It has a three different precipitators, and an on-line gamma power rating of 60 MW(t) and 19 MW(e). Construc- spectrometer with a Ge(Li) detector. The system tion began in May 1966, and sodium filling was com- immediately detected the two fuel element failures pleted by November 1969. The first criticality was of 1979 and 1980. Localization of a failed element achieved in August 1971, and the first electricity was is done in the following manner: fed into the public grid in August 1972. Because of problems encountered, operation was partially inter- Detection of different count rates of the rupted during the summer of 1973, and granting of delayed neutron signal in the two circuits, the power operating license was delayed. The prob- which allows the failed fuel element to be lems included sodium frost deposits in the gaps of the assigned to either the western or eastern rotating shield system and in the second shutdown sections of the cooler. rods; radiation resistant grease showed insufficient lifetime under certain mechanical loads in the main Comparison of the calculated and the valve drives; and an intermediate-size leak in the measured xenon 13l/xenon 134 ratios, steam generator caused a Na/H20 reaction. The leak which allows the distinction between a test was not strong enough to break the rupture disk in the element and a driver element. steam generator. In addition, a transient partial blockage of one fuel element remained unexplained. Behavior of the delayed neutron signal After December 1973, the KNK was operated accord- during skewed load operation, which ing to schedule. The operating and 100% power allows the failed fuel element to be license was achieved in February 1974.A-68-71 assigned to a core quadrant. Localization proper is then carried out by the dry sip- The transformation of a KNK into a fast reactor ping method in the refueling machine. during 1976 and 1977 necessitated changes to the reactor vessel’s internals, the fuel handling system, Disassembly and Postirradiation Examination-The and the reactivity control system. It also required a two failed KNK-I1 fuel elements for the SNR 300 new licensing procedure. This procedure posed spe- were dismantled in a hot cell without difficulty in a cial problems, as newer safety criteria were being reasonably short time. The postirradiation exami- applied to an existing installation; consequently, nation (PIE) of the two fuel rods involved in the the decay heat removal capabilities and the redun- 1979 fuel element failure indicated the defects were dancy of instrumentation cabling had to be not caused by design errors. Most probably, the n improved. defects were caused by fabrication techniques,

A-80 'c

Table A-8. Early power and test reactors

Reactot (Countrv)

DFR FERMI EBR-II* RAPSODIE** BOR-60 KNK-2 JOY0 FFTF PEC (UK) (USA) (USA) (France) (USSR) (Germanv) (Japan) (USA) (Italy)

General

Date critical 1959 1963 1963 1967/ I970 1969 I977 1977/198 1 I980 I985 Date full power 1963 1970 1965 1967/ 1970 I 973 1979 1979/ 1982 1980 - Electrical power, MW 15 65 20 - 12 21 - - - Thermal rating, MW 60 200 62.5 24/40 60 58 75/100 400 118

Core parameters

Core volume, L 120 400 73 49/42 60 250/70 304/238 1040 325 Core height, rn 0.53 0.77 0.36 0.34/0.32 0.40 0.60 0.60/0.55 0.91 0.65 Core diameter, m 0.53 0.7 0.51 0.43/0.4 1 0.43 0.82 0.7910.76 1.21 0.83 tp No. enrichment zones 1 1 1 I 1 2 1 2 I 22 Core volume fractions

%Fuel 40 29 32 40/37 40 32 33/35 35 35 %Sodium 40 50 49 34/36 28 50 41/41 41 38 %Other 20 21 19 26/27 32 18 26/24 24 21

Peak flux, I015n/cm2.s 2.5 4.5 2.5 2.0/3.2 3.3 1.U2.2 3.0/5.0 7.2 4.1 Ave. flux, 1015n/cm2.s - 2.6 1.5 1.212.3 - 1 .o 2.0l3.0 4.5 2.1 Peak linear power, kW/m 37 28 27 39/43 56 35/43 32/40 42 36 Ave. linear power, kW/m 35 17 23 26/3 I 35 16/31 - 24 25 Peak power density, kW/L 714 I002 650/1080 1180 - 480/701 730 512 Ave. power density, kW/L 900 458 860 4301770 750 150/440 279/391 460 350 A 4s 0. I4 0. I 0.1 1 /0.24 - 0.38 0.23/0.30 0.5 Peff 0.007 0.0068 0.005 0.006 0.005 /0.004 0.0032 0.0036 Doppler constant, -Tdk/dT - - 0.002 0.005 0.0037

Fuel assembly

No. driver assemblies 342a 102 127/77 64/54 Fuel type U-7% Mo alloy -b -C U02-PuO2 Enr. U02 UO~/UO~-PUO~UO2-PuO2 UO2-PuO2 UO2-PuO2 Fuel pellet form -d -e -e S A S S S A Fuel pellet density, %TD - IO0 96/92 93.5 90/86 94/93 90 95 Table A-8. (continued)

Reactor (Country)

DFR FERMI EBR-11* RAPSODIE** BOR-60 KNK-2 JOY0 FFTF PEC (UK) (USA) (USA) (France) (USSR) (Germany) (Japan) (USA) (Italy)

Smear density, &TD - - 85/75 89/84 73.5 83/80 87 86 88 Bond material Na -f Na He/He He He + Ar He He He Fuel pellet dia, mm 16.5 3.76 3.7/3.3 5.6/4.2 5.0 7.0/5.1 5.4/4.6 4.9 5.6 Pin diameter, mm 20.0 4.01 4.4 6.7/5.1 6.0 8.2/6 .O 6.3/5.5 5.8 6.7 Cladding thickness, mm - 0.13 0.23/0.30 0.45/0.37 0.3 0.5/0.38 0.35 0.38 0.42 Cladding material Nb Zr -g 316 SS1316 SS ss ss 316 SS 316 SS 316 SS Pin spacers Grid Wire Findwire Wire Grid Wire Wire Wire Pin pitch, mm 5.1 5.7 7.1/5.9 6.7 10.1/7.9 7.6/6.5 7.3 7.9 Pin pitch/diameter - 1.26 1.29 1.06/1.16 1.12 1.23/1.32 1.21/1.18 1.24 0.18 F.G. plenum location T None T T/T&B B T T B F.G. plenum length, m 0.06 NA 0.08/0.24 O.lh 0.40 0.41/0.55 0.9 0.50 Pin length, m - 0.83 0.46/0.61 0.4W0.53 1.1 1.36/1.55 1.91/1.53 2.38 I .65 No. pins per assembly 1 140 91 37/61 37 -I 91/127 217 91 Duct flat-to-flat, mm - -J 58.2 49.8/49.8 44 124 78.5 116 82.6 Duct thickness, mm 2.4 1 .o 1.0/1.0 1 2.6 1.9 3 2.4' Duct pitch, mm 68.4 58.9 50.8/50.8 45 129 81.5 120 86 Duct material SS ss ss/ss ss ss ss ss ss : Assembly length, m 2.45 2.33 1.66/1.66 1.58 2.09/2.30 2.97 3.7 3.0

Refueling

Refueling interval, days 60 14 315 35/90 - 365 60 100 60 Max. fuel burnup, MWd/kg - 10 80 50/90 127 80 50/60 80 33 Ave. fuel burnup, MWd/kg 30 8 68 - 80 50 42/50 45 40 Storage positions - 35 75 40/40 None - 20 -k 76 a. Number of rods. f. Metallurgical bond. k. Three modules of 19 positions each within reactor vessel. b. Enriched U-lO'f Mo alloy. g. 304LSS/316 SS. *MK-IA/MK-I1 Cores **Rapsodie/Rapsodie Fortissimo c. Uranium fissium alloy. h. 0.1 Top & 0.06 Bottom. Driver/Test Zone MK-I/MK-I1 Cores d. Annular fuel rods. i. 102-121/211-169. e. Fuel pins. j. 67.2 mm Square Duct. c

Table A-8. (continued)

Reactor

DFR FERMI EBR-11* RAPSODIE** BOR-60 KNK-2 JOY0 FFTF PEC (UK) (USA) (USA) (France) (USSR) (Germanv) (Japan) (USA) (Italy)

Reflectoriblanket

-a Material -b -g u02c U02-axial uo2iss Inconel uo2 SS-radial Axial

Material form Pins -b - P P P P Pins Material dia, mm 10.0 NA 13.3 5 .O 5.1 5.4/4.6 4.8 5.4 Pin diameter, mm 11.2 NA 14.5 6.0 6.0 6.3/5.5 5.8 5.8 Top length, m 0.43 0.35/0.39 0.27/- 0.10 0.20 0.40/0.30 0.14 0.18 Bottom length, m 0.43 0.16/0.57 0.27/0.24 0.10 0.20 0.40/0.35 0.14 0.25 No. pins per assembly 16 NA 7 + 7d/7 37 217/166 91/127 217 91 ;P 00 w Radial

No. assemblies -e 537 1621366 500/300 168 5 176/ 191 99 199 Material form R R -1 - - P Pk -f -f Material dia., mm 10.0 NA 15.4 - 8.0 13.6/- - - Pin diameter, mm 34 11.2 NA 16.5 14.5 9.2 15 .O/- - - Assembly length 2.33 2.45 2.33 1.12 - 2.14 2.97 3.7 3 .0 No. pins per assembly 1 25 6/19 7 7 121 19/- - -

Control

-P Material Enr.B4C -J Enr. B4C B4C and Et1203 Enr. B4C Enr. B4C B4C Enr.B4C No. assemblies 124 IO 10 6 7 8 6 9 I1

Reactor vessel

Height, m 6.3 11.1 2.28 9 6.2 10.2 9.9 13.3 10.7 Diameter, m 3.2 4.4 2.31 2.3 1.4 1.9 3.6 6.3 3.2 Thickness, m 0.012 0.05 0.019 0.015 0.018 0.012-0.016 0.025 0.019 0.030 Table A-8. (continued)

Reactor - (Country)

DFR FERMI EBR-II* RAPSODIE** BOR-60 KNK-2 JOY0 FFTF PEC (UK) (USA) (USA) (France) (USSR) (Germany) (Jauan) (USA) (Italv)

Containment structures

Configuration S S S S I cc S S D Materials I 1 3 .O 1 - 2/ 1 1 I 1/1 Des. press., MPa, gauge 0.137 0.22 0.17 0.24 - 0.25 0.13 0.067 0. I49 Des. leak rate, vol%/day 0.075 0.1 0.2 101 - 1 .o 3.0m 0.1 0.5

Heat transport and steam systems

Coolant NaK Na Na Na Na Na Na Na Na P Cover gas Ar Ar Ar Ar/He Ar Ar Ar Ar Ar 00 P Primary

Type Loop Loop Pool Loop Loop Loop Loop Loop Loop No. loops 24 3 1 2 2 3 2 Pump type E C C C C C C C C Pump position Cold Cold Cold Cold Cold Hot Cold Hot Cold Total flow, kg/s 450 I120 48 1 210/230 260 280 600 2200 630 Inlet reactor temp., "C 200 279 37 1 405/400 340 360 370 360" 400 Outlet reactor temp., "C 350 418 473 49515 10 520 525 468/500 503" 550 Max. fuel temp., "C 650 602 688 2000/2 I80 2055 2330/2500 2250" 2340 Max. clad. temp., "C 5 00 552 599 585/650 700 ~685 6201650 670" 650

Secondary

No. loops 12 3 1 2 2 2 2 3 2 Pump type E C E C C C C C C Pump position Cold Cold Cold Hot Cold Cold Cold Cold Cold Total flow, kg/s 900 1120 302 210/200 260 250 600 2200 624

Inter, heat exchanger

Number 24 3 I 2 4 2 2 3 L Primary side - Shell Shell Shell Shell - Shell Shell Shell c

Table A-8. (continued)

Reactor (Country)

DFR FERMI EBR-11* RAPSODIE** BOR-60 KNK-2 JOY0 FFTF PEC NJK) (USA) (USA) (Franc e) (USSR) (Germany) (Japan) (USA) (Italv)

Steam generator

Number 12 3 8 2 h Type - M M I/M - Tube configuration -h H S - - Superheater Yes Yes 2 Yes No Reheat er No Yes No No No

Turbine

Number - 1 1 - - 1 Inlet pressure, MPa 1 .o 3.97 8.62 - 8.82 7.85 ;P m Inlet temp., "C 270 404 435 - 460 485 ul -Dump heat exchanger Number 12 0 I 2 - 4 12

a. Natural uranium, Ni, & SS. f. Hexagonal Inconel blocks. j. Enr. U w/B4C follower.

b. Depleted U-2.75 Mo alloy: g. SS & depleted uranium. k. Three types (rod cluster, plate layer, block).

c. U02 below & SS above; UO2 radial. h. Tri-fluted block SS/plug w/flow holes SS. 1. at 0.025 MPa

d. Top + bottom. i. SS Hex blocks/pins. m. at 0.13 MPa 360°C.

e. 1872rods. which have since been changed. Friction marks water reactors. The fission product releases from such as material abrasion between the cladding the two failed fuel elements are relatively slight for tube and the spacers were not considered to be asso- the iodine and cesium remain in the sodium. The ciated with the defects. Disassembly of*the I9803 ..‘’,sodium’ represents an additional barrier to fission failed fuel element, which had been run with a fuel product releases to the environment. rod failure for about 4 weeks, revealed clear marks on the rods adjacent to the defective one and More precise analyses are conducted on the pri- seemed to indicate some interaction. However, mary sodium to detect the activation products detailed investigations have shown there had been sodium-24, sodium-22, Ag-1 10 m, and zinc-65, the no failure propagati0n.A-72 activated corrosion products manganese-54, cobalt-58, cobalt-60, iron-59, chromium-51, Thirty-five fuel rods from each of the two failed tantalum-182, antimony-124, the fission products, fuel elements were reprocessed in the Milli facility. cesium-1 37, cesium-1 34, antimony-125, barium/ The extracted fuel was used for fabrication of the lanthanum-140, and iodine-13 1. Because these fuel rods of the second core. Thus, in Germany, the radionuclides give rise to radiation exposure of per- breeder fuel cycle has been closed on a kilogram sonnel during repair and maintenance work, scale. In dissolving the fuel in 7 M nitric acid, some attempts are being made to remove these products, 10% remained undissolved and had to be retreated except sodium-24 and sodium-22, from the primary with hydrofluoric acid. The low solubility of the system. In KNK-11, investigations are concentrat- fuel is owing to its method of fabrication; highly ing on removal of manganese-54, zinc-65, and soluble fuel will be used in the first reload core. tantalum-1 82 isotopes. These nuclides have been shown to preferentially precipitate on specific metal Reactor Instruments- Performance capability was surfaces. Nickel, for example, lends itself particu- examined of temperature measurement probes and larly well for use as a radionuclide trap for zinc-65 two acoustic transducer probes, magnetostrictive and and mangane~e-54.~-~~ piezoelectric, which were inserted in KNK-I1 over a period of 4310 and 6635 h, respectively. These probes A stable evaporation zone is formed in LMFBR are planned for use in the safety system of the steam generators caused by the much smaller tem- German prototype LMFBR, SNR-300. The linearity perature gradients between the secondary sodium and temperature errors were demonstrated to be and the water vapor in comparison to the tempera- below 2%. Drift errors over a 4 week operating per- ture gradient of fossil-fueled heated plants. In addi- iod were less than 2.5% at sodium flow rates of less tion, there are no pulsating evaporation processes than or equal to 1 m/s. Drift errors were less than from local superheating. Impurities in the boiler 2%, with a sodium flow rate greater than 1 feedwater can become enriched in this stable evaporation zone and produce changes in the protective From the beginning, the reactivity signal of coating. An extensive measurement program is KNK-I1 showed harmonic oscillations of less than - being conducted at KNK-I1 in the steam water cir- 0.5 cents, Sensitive methods of correlation between cuit to determine what effects conductivity and the the reactivity signal and the measured fuel element concentration of oxygen, iron, copper, silicic acid, temperature proved these oscillations were caused and sodium have on this phenomenon. by individual fuel elements. Additional measurements performed under various operating condi- Measurements conducted at KNK-I1 while work- tions and theoretical considerations seem to ing without hydrazine have shown that hydrazine indicate the reactivity oscillations are initiated by cannot be used to reduce the oxygen concentration flow-induced mechanical vibrations. Most proba- to less than 50 ppm when there is increased oxygen bly, these are vibrations of entire fuel elements gen- content in steam in the presence of hydrazine. erated by vortex shedding or non-steady fluid jet Future investigations at KNK-I1 will be done with- formati0n.A-72 out hydrazine added but in a combined mode of operation involving an elevated oxygen content and Reactor Chemistry- Radial chemical measure- a reduced ph. ments of primary sodium show that the bulk gamma activity after 9,600 effective full power Experience with the Heat Transfer System - The hours of KNK-I1 operation was 1.9 mCi/g. This is sodium systems of KNK-I1 were started up in 1972. considerably below the primary coolant activation The primary system has since been continuously normally found in boiling water and pressurized filled with sodium, even during the KNK-I/KNK-I1

A-86 conversion phase. The sodium pumps have oper- 40 MW(t). The reactor provided design information ated uninterrupted and have accumulated some and fuel performance for use in France’s prototype 85,000 h. Systems, temperatures, and pressures reactor, PHENIX. have remained at the designed level for about 17,500 h. During this entire time, there have been Reactorfeatures-A view of the reactor assembly no major malfunctions. is shown in Figure A-59. The fuel assemblies, which contain both a fuel section and an axial blan- A dip tube has been built that extends vertically ket section, are in a central region surrounded by through the primary cell. In this dip tube a cali- radial blanket subassemblies. Removable subas- brated ionization chamber can be positioned at var- semblies outside this region serve as reflectors. The ious points. The ion chamber is used to measure the concentric steel cylinder assembly, as shown in Fig- radiation dose rate after the sodium-24 decays to ure A-60, is used for both a nonremovable reflector background values, about 10 d subsequent to reac- and thermal shield. tor shutdown. The residual radiation field is caused by activated corrosion products and fission prod- A jacket located outside the reactor vessel circu- ucts plated out on the walls of the primary system. lates gas for preheating. It also serves for cooling More than 10 years of operating experience has the reactor during shutdown. Steel containers filled shown the radiation exposure both of personnel with thermal insulation are located outside the pre- and of the environment is less than it is in compara- heating jacket. Additional structures consist of a ble water-cooled reactors. For instance, KNK-I1 containment tank and a biological shield. An outer was operated in 1982 using 44 man-rem with blast shield consists of three concentric layers: 43.9 man-rem acquired during maintenance on the borated graphite, lightweight cellular concrete, and primary system. Greater than 75% of the radiation reinforced concrete. Additional features are shown workers received less than 0.15 rem, and only 5% in Figure A-61. A single sodium inlet pipe is shown received greater than 0.4 rem. with two sodium outlet pipes located in ii vertical plane perpendicular to the plane of the injet pipe. The steam generators of KNK-I1 consist of parallel helical double tube systems. In 1972, a leakage in Core Arrangement-The fuel subassemblies are a steam generator caused by a weld pore was con- illustrated in Figure A-62. In the core regicln are the trolled safely and without any impacts on the envi- subassemblies, comprised of 37 fuel rods assem- ronment. One of the steam generators was removed bled in a stainless steel container with a hexagonal in early 1981 and replaced by a new one having a cross-section. Each individual element consists of a sampling point in the transition region between number of sintered pellets of mixed uranium and water and steam. This afforded an opportunity for plutonium oxide (25 wt% Pu02) contained in a metallagraphic reexamination. In the evaporator stainless steel tube. A spring arrangement main- region of the steam generator, a coating roughly tains contact between the various pellets in the 100 microns thick had developed; in the super- column, and provision is made in the upper region heater region, a coating roughly 40 microns thick of the element to collect fission gases. Separation had developed. The coatings consist of an inner between pins is maintained by a wire ~piral.~-~~ layer of oxidized metal and an outer layer of magnetite. The magnetite is generated whenever Both the upper and lower blanket sections con- the solubility of iron in water is exceeded. Pitting sist of a cluster of seven large-diameter elements of corrosion with a maximum penetration depth of depleted uranium. A lower base plate is provided to 250 microns is caused by the quality of water which distribute the coolant. The lower end of the subas- is not controlled in outage periods. The reduction semblies shaped in the form of a nozzle with an in wall thickness caused by pitting corresponds to internal venturi section reduces upward forces 8.6% of the wall thickness. induced by the coolant flow and aids in positioning. The upper end of the subassembly is provided France with a handling lug pierced with three holes for the sodium coolant outlet. The radial blanket subas- Rapsodie. The French LMFBR programs first semblies are similar to the fuel subassemblies, but major milestone occurred in 1967 with the commis- they contain a continuous seven-rod internal struc- sioning of Rapsodie. This was originally a 20-MW(t) ture of large-diameter depleted uranium elements. unit later increased to 24 MW(t) and then to Removable stainless steel reflector subassemblies

A-87 I _..

!- ..!- .._,r ..- ..,

Figure A-59. Vertical cross section of RAPSODIE reactor. are available to provide flexibility of operation and hexagonal guide tube similar to the hexagonal fuel fuel movement. assembly cans. The boron carbide rod is enclosed in a cylindrical can attached to a stellite nozzle and Control Elements-A poison-type control system handling head for the drive shaft gripper. Control is used that has four safety rods, each containing rod drive mechanisms are located in six sections about 100 g of B1o as boron carbide. In addition, around the cover plate equipment. Removal of the two regulating rods are provided, each containing control rods is accomplished by means similar to about 100 g of BIO. those used for refueling a cask car. Heat Transport system-Sodium is used as the Each of the six rods located on the boundary coolant in both the primary and secondary sys- between the core and the radial blanket in the high- tems. Heat rejection to air is planned for the sec- pressure coolant zone move a distance of 46 cm in a ondary system, rather than a steam generator

A-88 r Heating jacket Reactor vessel Low High Thermal shields pressure pressure

Reflector ~ I I I (non-removable) Zones

- Figure A-60. Horizontal cross section of RAPSODIE reactor.

. [SPECIAL CONCRETE REINFORCED CONCRETE ---. I

Figure A-61. Horizontal cross section of RAPSODIE reactor and shield.

A-89 Blanket rod -

+J Fuel pin

b1.96 in. 4

I wire Sect i0 n A-A

Figure A-62. RAPSODIE fuel subassembly.

A-90 turbogenerator installation. Two coolant loops are The tank acts as a second barrier in the event of provided, each designed for a 10 MW(t) load, util- vessel rupture. The containment tank is of plain izing centrifugal pumps. carbon steel 15 mm thick. Additional containment is provided by a radial biological shield consisting Sodium is pumped by a vertical centrifugal pump of borated graphite (3% boron) 600 mrri thick, rated at 97.2 L/s. Discharge from the pump tank which is maintained at a minimal temperature of enters the lower section of the reactor vessel and 220°C by cooling gas. then flows into the flow divider. Removal of the control rods is accomplished by means similar to Additional shielding includes a 0.4-m-thi ck wall those used for refueling a cask car. of cellular concrete, which protects the columns supporting the rotating plug. Also included is a Heat Tmnsport System Assembly-The flow baffle 0.4 m thickness of ordinary concrete, and finally a separates the coolant into a high-pressure zone, pro- 0.65 meter thickness of heavily reinforced concrete viding a coolant for the core and inner radial blanket maintained below 50°C. Thermal insulation is pro- as well as a low-pressure zone for the outer radial vided between the inner graphite and cellular con- blanket. The diffuser assembly provides a noncritical crete. A fixed plug is located between the geometry for the collection of fuel in the event of a containment tank and the reactor vessel, shown in meltdown accident. The coolant passes from the reac- Figure A-59. Internal to the reactor vessel are two tor blanket to the intermediate heat exchanger. Pri- rotating plugs: a large concentric rotating plug and mary sodium circulates through the shell at this a small eccentric rotating plug. Peneti-ations vertical stainless steel tube exchanger, and secondary through the rotating plugs are (a) a fuel handling sodium circulates through the tubes. In the design, port, (b) control rod drive shaft ports, (c) core the upper tube sheet is fixed, and the lower tube sheet cover plate drive shaft ports, (d) two ports for is floating to permit thermal expansion of the tube introducing experimental apparatus, and (e) two bundle. The tubes themselves are located in 12 con- ports for liquid level indicators. Two sets of ball centric sections. bearings are used for rotating the plugs. The liquid- metal seals of tin-bismuth provide a gas tig,ht seal Shielding Structures-Within the reactor vessel, between the exterior atmosphere and the reactor both the removable and nonremovable reflector cover gas. regions tend to reduce the neutron energy as well as attenuate the flux. The nonremovable portion of Provision is made in the design for a number of the reflector consists of 10 concentric cylindrical mechanical handling operations, for instance, the shields made of 24-mm-thick stainless steel and removal of fuel assemblies, control rods and their separated by an 8-mm space. Outside this assembly guide tubes, and the introduction of such items. A is a thermal shield consisting of four concentric cyl- handling port located in the small eccentric: rotat- inders having thicknesses of 4, 7, 10, and 10 mm, ing plug provides access for a gripper arrangement respectively. attached to a cask car that may reach any subassembly in the reactor. The cask car includes provi- The reactor vessel, shown in Figure A-59, of sions for cooling two irradiated subassemblies. Type 316 stainless steel, consists of a lower 12-mm- thick eliptical cylinder, a main 15-mm-thick plate, Transfer of subassemblies and other equipment and an upper section located between the fixed plug from the containment building is by means of a and a large rotating plug. secondary handling system, since the primary cask car is at all times within the building. A transfer Preheating of the reactor vessel is made possible compartment is provided for the movement of such by a 4-mm-thick jacket that permits circulation of subassemblies from one cask car to the other. Vari- an inert gas at 150°C. Preheating prior to the intro- ous other facilities are provided for necessary fuel duction of sodium minimizes thermal shock and handling operations. provides for the heating of sodium in the event of freezing. OperationalExperience- An extensive prog.ram of fuel evaluation and development has been carried Illustrated in Figure A-61 are steel containers out on the UO2-PuO2 mixed oxide fuel and on var- filled with thermal insulation or steel plates located ious uranium-plutonium-molybdenum ternary between the reactor vessel and a cylindrical con- alloys. The ternary alloy investigated consisted of tainment tank to prevent heat transfer by radiation. 75 wt% uranium, 15 wt% plutonium, and 10 wt%

A-9 1 molybdenum. Although the metal fuel is attractive outside of each fuel element; the ratio of the coolant n because of its good heat conductivity at the high flow is controlled by pressure drops. An upper axial power densities, the radiation experiments have blanket of natural uranium is included as part of each indicated excessive swelling. An additional disad- fuel element assembly. The core is surrounded radi- vantage of the alloys was their poor compatibility ally by a natural uranium blanket consisting of with stainless steel. A double clad design was neces- 2,000 e1ements.A-75 sary, consisting of an interior layer of niobium and an exterior layer of stainless steel with sodium After the EBR-I meltdown incident, the initial bonding. The cladding design prevented the forma- core design for Dounreay, which provided for bot- tion of a plutonium-iron eutectic. The oxide fuel tom support only, was revised to avoid the possibil- was chosen because it presents fewer fabrication ity of thermal bowing of the fuel elements. The problems than the alloy and because there is a core structure features a tube nest, shown in greater background of irradiation and production Figure A-64, which mechanically restrains the fuel experience with the oxide. elements at the top, center, and bottom. A series of long and short interlocking tubes are used with the United Kingdom long tubes located in counter-board sockets in the top and bottom plates. Side plates enclose the tube Dounreay Fast Reactor IDFRI. The Dounreay reac- nest, and the hole is positioned within the core skirt tor plant used a sodium-potassium eutectic, NaK, as a complete unit .A-74-75 as coolant. The plant was designed for a maximum power of 72 MW(t) and a core power of 60 MW(t). Details of the core fuel element are shown in Fig- The reactor design evolved from low power experi- ure A-65. Beta quenched uranium, enriched to ments performed at Harwell on the ZEUS experi- 45% and alloyed with a 0.5 atomic percent chro- ment. Construction was started in 1955, and the mium, was used as the initial fuel. The central ele- reactor went critical in November 1959. The reactor ment design was based on a maximum uranium was operated as a zero power assembly until April temperature of 840" C, a maximum coolant-inner 1960. It was then shut down for the installation of a vanadium cladding interface temperature of modified core tube nest to facilitate radiation test- 610°C, and a maximum coolant-outer neobium ing of fuel subassemblies for a prototype fast reac- cladding interface temperature of 5 15°C. The fuel tor.58959 An unusual feature of the reactor was the containers, or cans, are designed to restrict radial use of 24 coolant loops, each provided with a swelling and are provided with a free space to pump and heat exchanger to circulate the liquid accommodate axial expansion. Fission gases are metal through the reactor core and blanket. This permitted to escape into the coolant through a vent feature resulted from a design decision to use heat in the upper end of the fuel element.A-74-75 exchangers and pumps of a size with which experience had been accumulated in previous experimen- Controland Stability- Control is accomplished by tal loop work. movement of the fuel, a design decision based on the initial unavailability of BIO. Twelve groups of Reactor matures-The general arrangement of control rods, each consisting of 10 fuel elements, the Dounreay plant is shown in Figure A-63. A con- are situated around the end of the core. Six operate tainment sphere contains a reactor vessel and the as control rods, four are for shut-off, and two are primary coolant system. Provisions were made for safety rods. The reactivity involved is listed as shutdown heat removal by a natural convection 0.43 Ak/k for control, 0.028 Ak/k for shut-off, loop which vented up the tack.^-^^ The design and 0.014 Ak/k for safety. Interference with the was not optimized to represent a prototype. top plug arrangement and refueling systems is avoided by locating the control rod actuating unit Core Arrangement-The cylindrical core, 53.3 cm outside the rotating shields. Electromechanical in diameter by 53.3 cm high with a volume of 0.12 drives are used to avoid gas leakage problems. m3, contains 361 enriched uranium vented fuel ele- Movement of the control rod mechanism is ments in the form of hollow cylinders. These cylin- obtained by use of a vertical ball-nut and screw ders, with a 1.8-cm outside diameter and a 0.76-cm actuator driven through an electromagnetic clutch. inside diameter, are clad on the outside by niobium These features result in a relatively complicated and on the inside by vanadium. NaK coolant flows design. Reference A-76 presents a detailed descrip- downward, both through the center and around the tion of a safety control rod.

A-92 c

Figure A-63. Dounreay plant elevation. a -.-C L a c

Enlarged view at F

3 special tubes occuring at . center only F -Enlarged part plan on E-E

Section A-A

Section D-D

Figure A-64. Details of Dounreay core tube nest

Heat Transport system-Because the primary coolant leaves the bottom of the reactor vessel at objective of the.Dounreay reactor was to demon- 350"C, leaves the intermediate heat exchanger at strate the feasibility of fast reactor core operation, 200"C, then passes through the electromagnetic the heat transfer system was not designed as a suit- pump in a thermal siphon heat exchanger before reen- able prototype of an economical system. Instead, it tering the vessel. The pump was installed in the cooler was the most reliable combination of available part of the circuit because the coolant operating tem- components that would accomplish the purpose. perature was restricted at 200°C to prevent overheat- The coolant loops were constructed of piping varying at the windings. Each secondary loop removes ing in diameter from 1.9 to 15.2 cm without the use heat from two primary loops, necessitating 12 inde- of valves. Concentric tube construction was used pendent secondary units, each complete with a steam for the intermediate heat exchangers. These various generator. The intermediate heat exchangers, each restrictions led to the design requirement of 24 pri- designed for a 3 MW(t), consist of an inner stainless mary loops. A detailed discussion of the heat trans- steel pipe, 91 m long and 10 cm in diameter, sur- port system is given in Reference A-76. An unusual rounded by an outer stainless steel pipe of 15 cm feature of this heat transport design is the down- diameter which is formed into seven loops. Each of ward flow of coolant through the core. This per- the 12 secondary loops contains two electromagnetic mits various components at the top of the reactor to pumps installed in parallel in a cool return leg from operate in cooler regions and simplifies the struc- the steam generator. The steam generators, in turn, tural design of the core, since the coolant does not each have 6 MW(t) capacity and consist of 13 rows of tend to lift the fuel elements. 20 heat transfer elements each. These elements were constructed by spirally winding copper laminations onto stainless steel tubes. Because the tubes contain- The 24-loop primary system was designed to ing sodium are connected to the steam generating remove 60 MW(t) from the core and 12 MW(t) from tube by copper lamination, heat transfer is by conduc- the breeder blanket. As shown in Figure A-66, the tion from one tube to another via the copper. The n

A-94 t-48.35 t-48.35 in. overall length I

Core-bottom plate support plate

Niobium

Vanad i um Niobium S.S. bottom outerNiobium tube \ 4.S. top end end fitting 1 inner tube /adapter washer fitting

Enlarged view at X Enlarged view at Z Enlarged view at Y

Figure A-65. Details of Dounreay fuel subassembly.

Figure A-66. Arrangement of Dounreay primary circuit pipe work.

A-95 possibility of leakage from the NaK to the water sys- measurements by ion and fission chambers. An tem is therefore minimized. The heat transport system outer annulus of 5 wt% boron in graphite surround includes the usual array of accessory equipment; for these regions. The top shield includes layers of Q example, cold traps, hot traps, sampling equipment, borated graphite and steel plates in a rotating plug filling stations, and an inert gas (nitrogen) system. A and removable concrete block assembly. A 25-cm- thermal siphon system requiring no electrical power thick jabroc and steel shield covers the reactor top was installed to remove the shutdown heat of the core and is removable for shutdown operations. Second- from the primary circuits to a convection air-cooled ary shielding was provided by a 1.5-m-thick layer of heat exchanger outside the containment sphere. concrete on the inside of the containing sphere, which encloses the reactor and primary heat trans- Structural and Shielding Components- A view Of port system. the reactor vessel assembly with its associated components and shielding is shown in Figure A-67. The Operating Experience-A number of difficulties core and blanket, weighing approximately were encountered during the initial start-up proce- 90,000 kg, are supported by an inner structure dures of DFR.A-77-*4 Contamination of the within the vessel. This system, built up from a NaK with oxide proved much higher than antici- flange mounted on the inner skirt of the reactor pated and required considerable cleanup effort. vessel itself, includes the core skirt, various plates, The cold traps installed in each of the coolant loops and other members. In addition to providing the failed to operate effectively, and installation of a necessary support, the system separates the coolant better trap system and more satisfactory impurity flow to the core, inner blanket, and outer blanket measurement instrumentation were found neces- breeder. A system of rotating plugs forms the clo- sary. Starting in late 1961, there was considerable sure for the top of the vessel and permits access by stress corrosion in the 321 stainless steel superheat- the refueling machine to any fuel or blanket ele- ers and evaporators. The contributory causes were ment. Details are shown in Figure A-67. The plugs chlorides in the water, aerated conditions, and car- are removed only during shutdown. During normal bide precipitation, particularly at the surface of the operation, O-rings prevent leakage of the cover gas. scale. Extensive repairs were made to the steam During refueling, a leak-tight seal around the plugs generators, and careful control of water purity was is accomplished by two concentric metal rings dip- instituted. Repairs were not completed until April ping into a trough (a) of sodium-mercury amal- 1963. gam, in the case of the inner seal, and (b) of mercury in the case of the outer seal. Such a seal is Niobium and vanadium fuel cladding failures possible during refueling because the reactor pres- occurred as a result of hydrogen embrittlement sure is reduced to less than 3.5 kPa. When the ves- induced by hydride in the coolant. This contamina- sel is pressurized, the liquid metal in these seals is tion was believed to have been introduced with the discharged to a dump tank located on the plugs. first charge of fuel. Carbon in the system may also have contributed to the embrittlement of the clad- The reactor vessel and the primary coolant loop, ding. Considerable work was done to determine the which contains sodium-24 activated by a passage cause. Loops were in operation with a hot trap and through the reactor, is contained in a 41.2-m- specimens of zirconium, zircaloy, Nimonic, and diameter steel sphere, shown in Figure A-63. The three types of low-carbon stainless steel. Addi- primary coolant system and the intermediate heat tional efforts were made in the United States to exchanger are within the containment sphere (see determine the effects of carbon on structural mate- Figure A-66). Because the secondary NaK coolant rial. An extensive cold and hot trap system was is neither radioactive nor likely to contain fission installed to reduce oxygen, hydrogen, nitrogen, and products in the event of a fuel element failure, the carbon concentrations in the coolant to very low secondary loop system, including the steam genera- levels. Excessive niobium corrosion was reported as tors, is not within this sphere. Reference 58 late as March 1963. Considerable gas entrainment describes the shielding in detail. The design is based in the coolant circuits required a number of modifi- on a reactor output of 100 MW(t). Fast neutrons cations. These included changes in the expansion leaking from the vessel are thermalized in a 1.2-m- tanks and a separate gas enclosure for the control thick graphite shield containing 0.3 wt% of boron rod mechanism. which surround the reactor vessel. The graphite is cooled by recirculating nitrogen. Zones of pure Problems encountered with the control rod graphite in the thermal shield allow neutron flux thrust head were traced to the pickup of scum from n

A-96 I

Figure A-67. Cutaway view of the Dounreay reactor vessel and rotating shields.

the contaminated liquid metal surface. These prob- The new core tube nest, designed as Core B, was lems were eliminated by the hermetically sealed installed in April 1960. Many of the startup problems drive described above and in detail in Reference 76. were consequently solved, and a power level of During startup, problems were also encountered 11 MW(t) was reached in late 1961. The second fuel with numerous components, such as the control charge of 345 elements was discharged during the units and the mercury seals on the rotating shield. period December 21, 1961, to February 9, 1962. After the discharge, the primary system had four Low-power experiments were concerned with the complete fills of sodium and 12 dumps. Odly after determination of physics parameters. The predicted cleanup operations were completed in June 19152,and critical mass, neutron flux and power distribution for the holdup of oxide in the clogged bypass circuit elim- the core region, and the neutron energy spectrum at inated, was the plugging temperature reduced from the core center were reasonably well-confirmed. 230°C to about 140°C. The third fuel loading was

A-97 uranium-10 wt% molybdenum, and the loading was airborne contamination level during operation was completed in the period from June to July 1962. Dur- reduced to about eight percent of the maximum n ing this loading, the roller bearings on the charge permissible concentration. machine were replaced with ball bearings to prevent further bearing seizure previously experienced. Criti- During the early high power runs,' operators dis- cality was reached July 26, 1962, and operation at covered a slow but steady movement of gas from 30 MW(t) was realized August 7, 1962. During this the top of the reactor vessel to the outer gas ring period, the NaK rose in the rotating plug owing to main. Closing certain gas valves so as to divide the unbalance in the gas line, contaminating the mercury blanket into two separate sections confirmed the dip seal and requiring about 590 kg of mercury to free phenomenon. It was believed that fission product the seal. New balance lines were added to eliminate gas was accumulating above the core and diffusing this problem. Plugging temperatures were further into the ring main to produce the observed high reduced to 110°C while isolating the rotating plug radiation levels. Isolating the two sections of the graphite from the reactor cover gas. Testing of blanket system by retaining the more highly active uranium-1 8-at. %-molybdenum, uranium-20-at. 070- gas within the shielding effected a general reduction molybdenum, and uranium-25-at. %-molybdenum in radiation levels by a factor of ten. Regular com- indicated that the uranium-1 8-at. %-molybdenum moning of the sections was necessary to prevent the transformation from gamma to gamma + alpha buildup of excessive pressure differentials, which started in 40 to 45 min at 480"C, with completion of would lead to undesirable coolant level variations transformation in 150 to 250 h. Uranium-20-at. %- throughout the system. Further pipe work modifi- molybdenum at 430°C took 5 to 7 h to initiate the cations were made, resulting in a considerable delay same transformation and over 1,OOO h to complete it. being introduced into the time between gas collec- But at 550"C, it took 1 to 2 h to initiate and 200 to tion over the reactor and its appearance in the ring 300 h to complzte. The uranium-25-at.%- mains. Except for one or two highly localized areas molybdenum at 430 C took 50 to 70 h to initiate and into which access was normally denied during oper- was only 5% complete in 500 h. Examination of the ation, the net effect of these modifications and the uranium-molybdenum alloy fuels indicated fuel slugs extra shielding, which was installed on the ring irradiated to burnup at 0.3% had cracks that main, was to reduce the radiation to within accept- increased with fuel temperature, up to 450°C. Crack- able levels. ing was greatest in slugs that had the large fuel cladding clearances. Increase of molybdenum content from 20 at.% to 25 at.% showed how improvement Samples of the gas blanket indicate that the main in irradiation behavior had burned up 0.48 to 0.6%. constituents were xenon-1 33 and xenon-1 34- It was hypothesized that steep temperature gradients typical on-power values being lo8 and 6.2 x lo7 occurred as high as 790°C per cm in the fuel. Center Bq/L, and associated argon activity being 2 to swelling, produced by gas pressures in the high tem- 3 x lo5 Bq/L. Variation by factors of 100 were pos- perature center of the fuel, acted against the sible in these values, depending on the sample point restrained, cooler outer layers. The combined action chosen. Whenever changes were made in individual resulted in the cracks. primary pump loads, large variations in gas blanket radiation levels were observed, indicating that gas The need for shielding the exposed blanket gas was moving about the system. It became impracti- pipes had been very well established during 30-MW cal, without the installation of much more shield- 0perations.*-8~-~~It became evident that a very ing, to transfer the gas blanket to the storage tanks high standard of leak tightness was required of the immediately upon reactor shutdown. Normal shut- more than 80 joints associated with the instrumen- down procedures were modified to include a delay tation and gas blanket connections on the rotating of about 48 h before depressurization. shields. For example, during 30-MW operations in August 1962, the airborne activity level in the The coolant obviously contained considerable sphere was a factor of a hundred greater than back- quantities of primary fission products. The coolant ' ground, namely lo2 Bq/m3, and during the follow- monitors servey a small proportion of the primary ing run rose to lo4 Bq/m3, which exceeded the coolant for delayed neutron emitters, giving a high maximum permissible concentrations. Gamma continuous signal, which was calculated to be spectroscopy showed that the greatest component equivalent to the exposure of 4 x lo4 m3 of ura- of the activity was Rb-88, with the other major nium in the core (i.e., 25% of the fuel surface). The component being 0-138. Many of the joints were sensitivity of these monitors as indicators of fur- modified to include double rings. As a result, the ther fueI exposure, namely by element failure, was

A-98 therefore so small they were useless for this United States purpose. EBR-I/. The Experimental Breeder Reactor I1 The coolant had been regularly sampled at power (EBR-11), located at the Idaho National Erigineer- since the installation of the facility, which per- ing Laboratory, is rated at 62.5 MW(t), with an formed the sampling, and a typical gamma spec- electrical power output of 20 MW. EBR-I1 was trum determination was made in 1964 (described in originally designed to establish feasibility of Table A-9). metallic-fueled sodium-cooled breeder reaci ors for power plant service. It was also designed to demon- The amount of uranium and plutonium in the strate the feasibility of on-site fuel reprocessing primary circuit are very small, less than 0.5 g and techniques. By the mid-1960s all original ob.iectives 20 mg, respectively. The on-power coolant caused were met. As interest in fuel breeding developed by fission products activity is about 100 times that during the 1970s, EBR-I1 became the nation’s lead caused by sodium-24, with the dominant isotope facility for the irradiation testing of fuels, materi- being Cs-138 at 2.2 x 1O1O Bq/g. als, and instruments of interest to the breeder program. This testing was done from the viewpoint of After a fourth fuel loading, completed in June designing and constructing more advanced sys- 1963, the reactor operation reached 55 MW(t) on tems. Among EBR-I1 achievements are the follow- June 18, and a subsequent increase to 60 MW(t) in ing: generation of over 1.5 billion k’vlrh of July 1963. electricity; irradiation of over 10,000 specimens of fuel, structural, and absorber materials; on-site The 60 MW(t) run achieved a fuel burnup of testing of advanced instrumentation concepts; and 1.12%. Condition of the fuel cans externally was successful demonstration of an on-site, diversion- good. The fuel, as expected, was cracked and swol- proof system for fuel reprocessing. len. Operations through the fall of 1963 were at a maximum burnup of the uranium-molybdenum of The demonstration of on-site processing and 1.2 at.%. Electricity was first generated on fabrication is a procedure believed to have r1:sulted October 9, 1963, and a tie-in was made to the in favorable fuel cycle costs for high-power-density national grid on October 14, 1963. References fast reactors, and design of the entire reactor sys- A-77-80 cover operating experiments in Dounreay tem was affected by this requirement. A compara- in more detail up to August 1962. tively high enrichment necessitates the high power

Table A-9. Gamma spectrum six days after sampling

Energy Possible (MeV) Radionuclides Photons/min-g

0.14 Ce- 143/ 144 1.5 E+6

0.22 Te- 132 2.7E+6

0.364 1-131 1.52E+7

0.5 Ru- 103/ 106 1.5 E+6

0.67 1-132, Cs/Ba-l37/137m I .63 E + 7

0.76 Zr/Nb-95 5.18E+6

1.6 Ba/La-140 1.95 E+7

A-99 density, which in turn results in a high rate of fuel Reactor Core and Vessel Arrengement-Figure A-70 burnup. A finely divided fuel is necessary to shows the reactor vessel assembly; Figure A-71 shows achieve this high thermal performance. The fuel the core arrangeme~~t.~-~~In the radial direction, the achieves a high specific activity after irradiation, reactor is divided into three main zones: a core, an with the resulting limitations on fission product inner blanket, and an outer blanket. Twelve control separation processes. A high level of fission prod- rods are located at the outer edge of the core, and two uct decay heating also results; thus, there is a need safety rods are located within the core. The core, for substantial cooling in the reactor during fuel including the control and safety rods, have an equiva- handling and fuel cycle. The fuel also has a high lent radius of 21.17 cm and a height of 36.12 cm, monetary value, which provides incentive to mini- with a total core volume of 0.07 m3. Located in the mize total fuel inventory by reducing out-of-the- core zone are 47 core subassemblies, each containing reactor processing time requirements. 91 fuel pins, two core assemblies acting as safety rods, and 12 core subassemblies acting as control The fuel cycle involves pyrometallurgical proc- rods. In addition to the central core section, each core essing, which can be accomplished with relatively subassembly contains an upper and lower blanket sec- small equipment demands and a short cooling tion. The 12 control rods and the two safety rods con- time. Incomplete decontamination, however, asso- sist of modified movable core subassemblies. These ciated with the process necessitates the use of rods are moved in stationary thimbles having external remote control procedures for the fabrication of dimensions and lattice spacing identical to the core new fuel elements from the processed product. The and blanket subassemblies. fuel process itself has likewise affected the core design, specifically the composition of the metallic In each radial zone of the reactor, the hexagonal uranium alloy for the first core. The fuel composi- subassemblies of identical size measure 5.82 cm tion is a metallic alloy, 95 wt% uranium enriched to across the external flats and have a heavy wall thick- 65% uranium-235, and 5 wt% fissium. Fissium is ness of 0.1 cm. Each subassembly contains a number an equilibrium concentration of fission product of fuel and blanket elements with size and shape elements left by the pyrometallurgical reprocessing appropriate to the particular type of subassembly. cycle designed for EBR-11. It consists of 2.4 wt% Mo, 1.9 wt% Ru, 0.3 wt% Rh and 0.2 wt% Pd, The subassemblies are spaced on a triangular 0.1 wt% Zr, and 0.01 wt% Nb. Fissium contrib- pitch with a 5.89-cm center-to-center distance. This utes desirable characteristics to the fuel. Both reac- provides a nominal clearance of 0.076 cm between tor and the associated fuel recycle facility were each subassembly to permit removal of the units designed to provide a highly flexible installation from the reactor. Each face of a core or inner blan- that would permit investigation and evaluation of ket subassembly hexagonal tube contains a projec- various core configurations, types of fuel, fuel ele- tion or button, 0.95 cm in diameter by 0.036 cm ment design, and processing techniques.A-21 y30- high, located approximately at the horizontal cen- 35,81,86-91 ter line of the reactor, providing a plane of contact at that location. All subassemblies, including those Re8ctor Fe8tures-Figures A-68-70 show the for control and safety, have an identical upper major reactor component^.^-^^ The fuel cycle facil- design and are accommodated by the same han- ity is seen to be a major portion of the installation. dling and transfer devices.

The reactor and the primary heat removal sys- The core subassembly consists of three active sec- tem, including the intermediate heat exchanger, are tions: an upper blanket, a core section, and a completely submerged in a large bulk volume lower blanket. The fuel elements in the core's cen- (300 m3) of sodium within the primary tank. Fuel tral section are pin type and consist of a right circu- handling is carried out under sodium. lar cylinder of fuel alloy fitted into a thin-walled stainless steel tube. A 0.015-cm sodium-filled A steel containment vessel is used to house the'pri- annulus between the pin and the inside of the tube mary system components, and power generation provides a thermal bond. Above the sodium, an facilities are housed in a separate building. Approxi- inert gas space is provided to accommodate sodium mately 85% of the power is generated in the core, with expansion. The 91 cylindrical fuel elements (pins) a power density of 890 kW(t)/L and an average spe- contained in the core section are spaced on a trian- cific power of 314 kW(t)/kg of uranium-235. gular lattice by a single helical rib on the outside of

A- 100 Cover4 ift i ng me(: han is m

Subassembly gripper mechanism S torage-rac k mechanism Control mechanisms

Subassembly transfer mechanism

Shield-cooling duct

Primary tank cove

'Mechanical pump

Reactor vessel

Figure A-68. EBR-I1 major reactor components and reactor primary tank.

each element. The lower ends of the pin are fas- lar lattice. The pins are unalloyed depleted ura- tened to a parallel strip support grid with the upper nium, 0.8 cm in diameter, and 45.7 cm long. They ends unrestrained. are contained within a stainless steel tube having a 0.02-cm sodium-filled annulus to provide the nec- The identical upper and lower blanket sections essary thermal bond. Bond ends of the blanket ele- consist of 19 pin type elements spaced on a triangu- ments are positioned in the subassembly by a /\

A-101 n

7Control rod drlve shafts Reactor vessel cover - torque pins Reactor vessel cover lock mechanism

Reactor vessel cover

Thermal baffle Flow baffle -1 Inner blanket subassemblies Outlet plenum Outer blanket subassemblies

Reactor vessel -Core subassemblies

Outer neutron shield liner Inner neutron shield cans Inner neutron shield retainers Reactor liner- Outer neutorn shield cans Upper grid plate High-pressure coolant plenum

Note: For plan view see Figure 4. Figure A-70. EBR-I1 reactor vessel assembly.

A- 102 Outer-blanket section W Retaining shell In ner- blan ket sect ion

Neutron- r Core section

Thermal baffle

Reactor vessel

Neutron shield

Shieldi ng-can spacer

Inner retaining shell ’ Outer retaining shell \ Figure A-71. EBR-I1 primary neutron shield.

parallel strip grid similar to that used in the field sembly only in the design of the lower adapter. In sections. Axial expansion is permitted, but other the case of the outer blanket subassembly, a small- movements are restricted. The upper adapter of the diameter adapter contains an opening at the bot- assembly is provided with an attachment knob for tom through which the coolant enters. the various fuel handling gripper units and a collar for the transfer arm. The lower adapter is a cylin- Control Rods-Operational control of thc reac- drical inlet nozzle for the coolant, and also serves tor is achieved by 12 control assemblies with a fuel as a locating and support plug in the reactor grid pin core section identical to that in other core posi- plate. The bottom end of the nozzle is closed; the tions. Control is by removal or insertion of the fuel. coolant enters the nozzle through holes in the cylin- Instead of an axial blanket, a void section normally drical wall. filled with sodium is above the core section, and a steel cylindrical tube reflector is below the fuel sec- The inner and outer radial blanket subassemblies tion. The middle section of the rod consists of a each consist of 19 cylindrical blanket elements modified core subassembly comprising 61 fulel ele- spaced on a closely packed triangular pitch and ments identical to those used in the other core sub- contained in a hexagonal can. The central active assembly units. This assembly is encased in a blanket section consists of depleted uranium cylin- 4.85-cm-across hexagonal tube, which is smaller ders 1.1 cm in diameter and 1.4 m long. They are than the normal hexagonal tube by the equivalent contained in a stainless steel tube with a 0.03-cm of one row of fuel elements. sodium annulus as a thermal bond and an argon gas expansion region above the sodium. The enclo- The drive arrangement for the control and safety sure unit is sealed with welds. The two types of rods is shown in Figures A-72 and -73. The two blanket subassemblies differ from the core subas- identical safety rods employed to provide shutdown

A- 103 Limit Shock n

Rod-position synchro transmitter

Common drive unit -

Drive shafts Safety rods j_

Figure A-72. EBR-I1 safety rod drive system.

reactivity during reactor loading operations are seat disconnects and the subassembly holddown driven from below the core. They are essentially devices at the bottom of the vessel. The coolant identical to the control rods except for modifica- passes into a common outlet plenum chamber and tions at the lower end which provide the necessary then out of the reactor through a single outlet noz- attachment to the drives. zle to the intermediate heat exchanger. After passing through the heat exchanger, the coolant returns Heat Transport Systems-A unique feature of the to the bulk sodium supply at approximately 370°C. heat transport system is the primary tank. It con- This temperature is slightly above the bulk temper- tains the reactor vessel and primary coolant system ature, since it is necessary to compensate for the submerged in a large volume of bulk sodium, as small heat losses from the primary system tank. For shown in Figure A-68. maintenance, the pumps may be removed from the primary tank by ball seat pipe disconnects. The Two identical vertically mounted single-stage heat exchanger shell is permanently attached to the centrifugal pumps, operating in parallel, force cover of the primary tank, but the tube bundle and coolant from the bulk sodium in the primary tank associated structure are removable as a unit. Addi- through the reactor at a rate of about 536 L/s. tional component details are given in Reference A-93. Flow through the reactor itself is 517 L/s. Of that flow, 440 L/s move from a high-pressure plenum In the secondary system, an A-C electromagnetic through the core and inner blanket subassemblies; linear induction pump, which provides close con- 44 L/s flow from a separate low pressure plenum trol of the flow rate, circulates sodium through the through the outer blanket subassembly; and 32 L/s tubes of the primary heat exchanger at a flow rate flow through the clearance spaces between subas- of 780 L/s. The sodium then enters the super- semblies. The remaining 19 L/s represent leakage heated section of the steam generator at 470°C and back to the primary tank through the pump ball returns to the pump at 320°C. n

A- 104

I. Ls R Control-rod drive mechanism

Safety-rod drive mechanism

12 Ift-4I in. I 'Main floor

Control-device I ifti ng mechanism

Rotating-shield plugs

E Sodium level

Guide bearing

>Drive shaft 13 ft-9-112 in. / Neutron-shield cover / Upper plenum chamber /Gripper and sensing device

1 Safety-rod guide thimble ------Neutron shield Reactor core 7 ft:7 in. /Support grid Rod drive beam

3 ft-4-112I in. -/Bottom of primary tank LLc,

Figure A-73. EBR-I1 control and safety rod drive systems.

A-105 The steam generator, shown in Figure A-74, is the secondary system. Feedwater heating is accom- n one of the natural recirculation types with separate plished by extraction from the main turbine, by superheaters and evaporator sections connected to exhaust from the feedwater pump turbine, and by a single steam drum. Each evaporator unit is con- high-pressure steam from the main steam line. nected to the steam drum by a single downcomer and riser. Dry and saturated steam from the steam Structures and Shielding Components- In EBR-11, drum passes down through the superheater unit. the reactor vessel assembly is completely sub- Double-walled tube construction is used for both merged within a large volume of bulk sodium con- the evaporator and superheater units; no welds are tained in the primary tank. The assembly itself is used in the portion in contact with the ~odium.~-~~mounted on the bottom structure of the tank in such a manner that the distance from the top of the At full load [62.5 MW(t)], when the generator reactor vessel to the free surface of the bulk sodium produces 20,700 kW electrical gross, the steam sys- is approximately 3.6 m. The reactor vessel assem- tem flow includes 0.63 kg/s of steam bypass to the bly, as shown in Figure A-70, includes the lower condenser. A full-capacity automatic steam bypass grid plenum with inlet nozzles and safety-rod system permits reactor operation without turbine drives, an inner radial neutron shield, the subas- generator operation or with turbine load at any sembly holddown and flow-baffle structure with fraction of reactor power. This bypass system pre- outlet plenum and outlet nozzles, the vessel cover vents major load changes from effecting changes in and hold down structure, thermal baffles, the

- Sodium inlet 1

Figure A-74. EBR-I1 steam generator.

A- 106 cylindrical vessel wall, and the outer radial neutron system consists of a gripper mechanism, a subas- shield .A-68992 sembly hold down rod, and transfer arm assembly. Positioning of the gripper mechanism is itccom- Located within the primary tank, but outside the plished by rotating two eccentric plugs in the shield. reactor vessel, are the fuel handling components, Details of the fuel handling procedures and equip- which are an important feature of the reactor ment are given in Reference A-87. design. The system, shown in Figure 75, has the capacity to remove a subassembly from the reactor, The shield system consists of two sections. A transfer it to the storage rack, and remove it from neutron shield between the blanket and the bulk the primary tank after a 14-d cooling period. The sodium tank, reduces neutron leakage into the

Gripper mechanism

Fuel-unloading machi ne ', I

Sodium level

Reactor cover (raised position)

Reactor cover hold-down Control drive shafts

Subassembly hold-down

Figure A-75. EBR-I1 principal fuel handling components

A- 107 sodium and thus reduces activation. A biological The primary heat exchanger, located within the n shield outside the sodium tank reduces radiation primary tank, is surrounded by a 2.5-cm-thick shell levels to acceptable limits at accessible locations of 1.5% borated steel to reduce activation of the around the reactor. This arrangement is shown in secondary sodium system. Figures A-76 and A-77. Within the reactor vessel, part of the neutron shield consists of two rows of The biological shield, outside the primary tank, stainless steel cans, each 10.2 cm square, filled with consists of a 1.8-m thickness of ordinary concrete graphite. Five additional rows are located outside in the radial direction wherever space is available. the reactor vessel; the fifth and seventh rows consist Also making up the biological shield is a top shield of 3% boron carbide. Graphite is substituted for consisting of a layer of steel balls directly above the the boron carbide in areas immediately adjacent to primary tank. Heavy concrete is used elsewhere. the neutron detecting instrument^.^-^^ Rotating plugs in the top shield are stepped to

1. 54017 crane 2. 75-ton crane 3. Crane bridge 4. Concrete missile shield 5. Gripper hold-down mech. 6. Control-rod drives 7. Storage-rack drive 8. Rotating plugs 9. Blast shield 10. Primary coolant auxiliary pump 11. Reactor-vessel cover 12. Neutron shield 13. Basement 14. Sodium purification cell 15. Na-to-Na heat exchanger 16. Reactor 17. Subassembly storage rack 18. Concrete biological shielding 19. Subbasement 20. Primary tank

. El. 102 ft-0 in.

El. 89

' El. 87 ft-4 in.

El Figure A-76. Vertical cross section of EBR-I1 reactor building. 9 A- 108 Reinforcing bars \

Celotex J Concrete biological shield

Figure A-77. EBR-I1 blast shield and biological shield.

reduce streaming through the large number of The Instrumented Subassembly Test 1:acility voids and penetrations necessary to provide access. (INSAT) is used primarily for the irradiation of fuel Control rods, however, are not stepped, but stream- materials. The In-Core Test Facility (INCOT) is ing is reduced by keeping clearances as small as pos- used for irradiation of structural materials, absorb- sible and providing more thickness than otherwise ers, and sensors. Both differ substantially in design required. Additional details of the shield system are and operation. given in Reference A-92. A steel containment vessel 24 m in diameter and approximately 42.7 m high The design of INSAT was constrained by two houses the reactor plant. principal considerations: (a) the need 1.0 run instrument leads through a control rod penetration Instrumented In-core hciIity-The needs for gather- and the reactor vessel cover, (b) the need to b:ep the ing physical data from subassemblies under irradia- discharged fuel bundle under the primary $;odium tion were not considered in the original design of at all times during its removal from the core These EBR-11. However, as the irradiation program acceler- constraints were eventually satisfied by replacing a ated during the late 1960s, increased interest was control rod (and its associated drive) by a single expressed in an in-core irradiation facility. Such a unit that can be loaded from the reactor floor into facility would permit cabling connections between the reactor grid, and which permits cable connec- sensors in an experimental fuel bundle and the out-of- tion between the in-core sensors and out-of-core core recording instruments. In the late 1960s and recorders. The unit is a fuel bundle physically simi- early 1970s, two instrumented in-core irradiation lar to a control rod and a double-walled exlension facilities were added to provide flexibility to the irradi- tube, and is mechanically latched to the tlundle. ation program, and to permit continuous monitoring Cables are led from the sensors up to the ciouble- of physical data for specimens being irradiated under walled tube and into a terminal box located outside typical operating conditions. the primary tank. During routine operation, the

A- 109 fuel bundle occupies the same position as that nor- tion of the container is remotely controlled from a mally occupied by a control rod. During fuel han- shielded booth. As for INSAT facilities, the entire dling operations, the entire in-core assembly is assembly is lifted prior to fuel handling operations. raised to an elevation that permits normal functioning of fuel handling equipment. Following fuel Since 1972, the INCOT facility has been used to handling operations, the assembly is lowered back irradiate 10 principal series of instrumented experi- into the core for subsequent irradiation. ments. Among these were the irradiation of absorber materials (B4C and Eu-02), self-powered When test bundles reach their target burnup and neutron detectors, eddy current flow sensors, are ready for removal from the core, a special cutting acoustical monitors, and biaxial creep specimens. tool is lowered through the extension tube. Rotation of the tool cuts the instrument leads. The extension A more recent in-core instrumented facility, the tube and cutting tool are then drawn up into a pulling Breached Fuel Test Facility (BFTF), has been pipe, leaving the fuel bundle ready for transfer with installed. This facility was designed to test the fis- the standard fuel handling equipment. sion product characteristics of breached fuel under normal operating conditions. BFTF consists of a Depending on the needs of the experimenter, as few device that monitors the exit sodium from a test as one and as many as 61 other specimens may be vehicle located immediately underneath. The test accommodated in the test vehicle. As of vehicle, in turn, can contain any one or more of a December 31, 1982, seven individual sets of instru- variety of experimental fuel elements, either intact mented fuel specimens had been irradiated in the or intentionally breached. BFTF is equipped with INSAT facility. Typical information derived from the two flowmeters, a delayed neutron detector, and a tests include temperature, coolant flow rate, gas pres- deposition sodium sampler used to study the plate- sure, creep rate, and neutron flux. out of fission products.

Test specimens in the INCOT facility may be Another in-core instrumented facility, the Fuel inserted or removed from the reactor floor area. Performance Test Facility (FPTF), has been The distinction between the two facilities is one of installed. Principal features of FPTF include the use. Because of decay heat considerations, fueled following: a throttle valve for changing and con- specimens must be kept submerged in the primary trolling the coolant flow rate through the facility, a sodium following irradiation. Such a constraint is sodium flowmeter, thermocouples, and a delayed more easily satisfied by treating the test section dur- neutron detector. Because its coolant flow rate can ing its discharge as a spent fuel assembly. Further- be changed during operation, FPTF can be used to more, attempts to discharge irradiated fuel cycle temperatures in fuel element specimens. specimens via the reactor floor would pose major FPTF can, of course, be used to study the effects of shielding problems. transient operation in both intact and intentionally breached fuel elements. For nonfuel test specimens, decay heat problems are essentially nonexistent. Associated radiation At the nuclear Instrument Test Facility (NITF), fields are such that shielding requirements during capability exists for on-site performance testing of discharge through the reactor top can be satisfied. nuclear instruments under LMFBR conditions. Two of eight existing instrument thimbles are used The INCOT facility is, in its simplest form, a for this purpose. In the original configuration, thimble assembly that extends upward from the EBR-I1 was fitted with four J-type nuclear instru- reactor grid (below the core) through the core, reac- ment thimbles that extend down from the upper tor vessel cover, and small rotating plug, to the reac- shielding to the primary tank and into the neutron tor floor level. Instrument leads run directly from shield around the vessel. Four other 0-type thim- test specimens or sensors up the thimble to a top bles similarly extend down from the upper shielding mounted terminal box. Special handling systems to terminate outside the neutron shield. One permit removal of irradiated test bundles from the 0 thimble and one J thimble have been converted facility, reinsertion of reworked irradiated test bun- to instrumented test facilities. dles into the facility, or removal of the facility itself. Removal operations involve the withdrawal of the The NITF thimbles are 8.5 m long and 38 cm in test bundle into a handling container 11.3 m long, diameter at the lower ends. Instruments are usually which is suspended from the building crane. Opera- located at core mid-plane elevation. Leads from the

A-1 10 instruments are led up through the thimble to the operation consisted mainly of slitting the wrapper reactor operating floor. can and removing individual elements from the fuel bundle. Fuel elements, in turn, were loaded into Test temperatures ranging from 66 to 370°C are magazines and moved to the argon cell via im inter- made possible by varying the amount of cooling air connecting lock. circulating through the facility. If higher temperatures are needed, thermostatically controlled ovens Initial operations in the argon cell consisted of may be used for instrument heating. Data taken the following: shearing upper and lower ends from under operating conditions allow the experimenter the fuel elements, removing spacer wires, removing to evaluate effects of high temperature and intense the cladding, and chopping the spent fuel pins into radiation fields on the performance of various neu- 3.5-cm lengths. Chopped fuel in amounts (of 10 to tron detectors and instrument cables. 12 kg, along with precalculated amounts of enriched uranium, were melted in CaO-coated zir- The Radioactive Sodium Chemistry Loop conia crucibles and held for 3 h at a temperature of (RSCL) is a facility that permits development test- %I ,400”C. During this period, approximatdy two- ing of techniques to measure impurity levels and thirds of the fission products-those wiirh high primary sodium. The facility consists of five vapor pressure, such as xenon, krypton, iodine, shielded cells, a 5. I-cm-diameter main loop, and bromine, cesium, and cadmium-left the melt smaller branch lines that deliver primary sodium to through volatilization. Less volatile electropositive and from the various cells. Coolant is pumped fission products, such as yttrium, barium, stron- through the main loop by a dc electromagnetic tium, and the rare earths, reacted with the i:irconia pump at flow rates and pressure variable up to to form an oxide that remained within the crucible 1.9 L/s and 0.11 MPa, respectively. after pouring operations. The volatized fission products were collected by a Fiberfax cover over the Each cell may be isolated from the main loop, crucible, which also collected sodium used as a and, after a suitable decay period, physical access is thermal bond in the fuel. The purified alloy was permitted for the installation and maintenance of poured into the mold to provide the ingot for fabri- equipment. Extensive precautions have been taken cation; the oxidized fission products, principally to prevent, annunciate, and minimize the impact of the rare earths, were left behind in a sci-called sodium leakage throughout the entire facility. “skull” retained by the crucible.

The facility has been used extensively as a test The pouring operations were never totally effi- bed for proving the application of various on-line cient. Inevitably, a relatively small but corisistent impurity measuring devices. Among devices tested amount of fuel, .~6-8%,remained in the crucible in the RSCL are a plugging temperature indicator, with the dross. These skulls were set aside in hot tritium monitor, vacuum distillation sampler, oxy- storage for subsequent uranium recovery. gen and hydrogen meters, segregated iodine sampler, graphite-cesium trap, and an equilibrium The process contemplated for recovering the module for carbon analysis. skull material is to convert it to an oxide powder suspended in a molten chloride flux, the noble Fuel cycle--Because a very important objective metals to be removed by reduction and extraction of this reactor is the demonstration of an integral by zinc. A dilute magnesium-zinc alloy wouid then fuel processing facility, some features associated reduce the uranium oxides and other fissiori prod- with the fuel reprocessing are given here and are uct oxides. A uranium zinc intermetallic compound found in more detail in References 94-109. The would be precipitated from this metallic sohtion by first fissium core consisted of stainless steel jack- cooling, and the desired uranium recovered in sev- eted pins containing 50% enriched uranium alloyed eral additional reprocessing steps. Although the with 5% noble-metal fission product elements. melt-refining separation itself is quite simple, the Irradiated fuel assemblies were allowed to cool 14 d need for the skull recovery process may result in a in the primary tank and were then removed to the complicated overall operation. Fuel Cycle Facility (FCF). Transfer operations were relatively simple. Spent fuel subassemblies were The first step in fuel pin casting operatioris con- placed in the air cell for disassembly operations, sisted of remelting a melt-refined ingot in a thoria- using a system of shielded casks, airlocks (between coated graphite crucible with an induction heated the reactor building and the FCF), and cranes. The vacuum furnace. Also located in the furnace in a

A-1 11 vertical attitude above the crucible was a cluster of scheduled intervals and the need to replace spent n approximately 100 Vycor molds. After the charge fuel with fresh fuel. Approximately 7 d are needed was melted, the crucible was raised to a position between runs to accommodate refueling operations that immersed the lower end of the mold cluster to and perform minor maintenance activities that a depth of 3.8 cm. The furnace was then rapidly cannot be carried out with the plant running. Each pressurized to 0.17 MPa to drive the melt up into year the plant is shut down for four to six weeks to the evacuated molds. After a few seconds, the melt carry out comprehensive modification, mainte- froze and the crucible was lowered to its original nance, and inspection. position. Following a 4-h programmed cooling period, the furnace was opened and the molds were The operation of EBR-I1 has always been kineti- removed. Unused fuel material in the crucible was cally stable. A prompt negative power coefficient broken up and returned to the fuel stream. of reactivity from the expansion of coolant and fuel effectively damp the effects of small reactivity The first step in fuel pin processing operations changes from inlet temperature variations and con- consisted of breaking the Vycor molds away from trol rod motion. The amplitude of the prompt the castings with a pneumatically actuated crushing power coefficient component is occasionally mea- system. Pins, collected in trays, were fed by gravity sured by rapidly withdrawing reactivity from the to a shear that cropped the pins to required lengths. system, dropping a special rod, and analyzing the Following cropping, the pins were subjected to non- shape of the power decay curve. destructive tests that consisted principally of measurements for weight, length, diameter, and During the period from 1976 to 1982, EBR-I1 porosity. Rejects were fed back to the melt refining operated with plant availability factors greater than operation. 70%. A peak value of 77.1% was reached during 1980. Plant availability factors in this range com- Acceptable pins were loaded into stainless steel pare favorably with those of commercial nuclear tubes containing solid sodium wire (%0.65 to and conventional fossil fuel power plants. If down 0.85 g). The sodium was melted and the fuel pins time required by the experimental programs is dis- settled by gravity. End-plugs were inserted into the counted, the actual plant availability factor during jackets and peripherally welded to the jackets with the 1976 to 1982 period exceeded 80%. a remotely operated capacitor-discharge welder. To ensure void free sodium bonds, the finished ele- Refueling time between runs is not a serious con- ments, clustered 50 to a magazine, were heated in a straint. The capability of interim in-tank storage furnace 1 h at 500°C. The elements were then sub- permits the pre-shutdown transfer of fresh fuel jected to %l,OOO vertical impacts (under 500°C assemblies to the storage basket. Approximately conditions) at the rate of 100 impacts per min. The 4 h after shutdown, spent subassemblies may be finished elements were then inspected for bond, transferred from the core to a storage basket and flaws, and sodium level. The final step consisted of replaced with fresh subassemblies. The turnaround monitoring the elements on a grid system and time per subassembly amounts to approximately incorporating the resultant fuel bundle in a fresh 1 h. In the absence of problems, the time required hexagonal wrapper can for subsequent return to the for end of run refueling operations amounts to reactor. approximately 24 to 48 h. The interim fuel storage feature is beneficial in another important respect. After nearly 5 years of successful operation, the After fulfilling minimum cooling requirements, FCF was shut down in 1969. Subsequent core changes spent subassemblies may be transferred into and have been made, either by vendors in the commercial out of storage baskets while the reactor is running. sector or by on-site personnel in a companion “cold” line facility. Currently, the argon cell is on standby Routine operations are occasionally interrupted status. The cell has been stripped of equipment and is by the release of gaseous fission products from a in the process of decontamination. failed fuel element. The majority of the releases are the result of endurance tests in which fuel elements Operating Experience-Operating run lengths of are intentionally irradiated to failure and beyond. EBR-I1 are nominally 2,800 MWd or 40full- On other occasions, the failure may be the result of power d. Two factors determine the run length: the a faulty weld or a premature failure of cladding. need to discharge the radiation experiments at More recently, the effects of sustained operation n

A-112 ..

under breached cladding conditions have been lyzed for a wide variety of fission product and cor- investigated. Fuel elements that failed under irradi- rosion species. ation were permitted to remain in the core for periods up to a few weeks to evaluate the effects of Many years of operating experience have demon- sustained operation on the fuel element. Also eval- strated the feasibility of conducting routine and uated were the consequences of releasing fission specific maintenance on LMFBR systems and com- products for the primary coolant and cover gas ponents. Considerable maintenance experience system. with unconventional, relatively accessible, and moderately radioactive components has been A portion of the EBR-I1 irradiation program is achieved. A typical component falling into 1 his cat- concerned with effects of 'fission products on the egory is the transfer arm (part of the fuel handling primary sodium and cover gas systems. Species system). Precautionary measures associated with related to the coolant are removed, in part, by such an activity is to maintain an argon atmosphere means of an externally located cold trap. Gaseous around the component to prevent any leakage of air fission products released to the cover gas system are to the cover gas system. The measure also prevents removed by a combination of cryogenic trapping leakage of cover gas to the reactor buildin,g. Such and absorption using liquid nitrogen-cooled char- an activity is based on pulling the component into a coal beds. rubberized nylon bag sealed by a flange to a penetration in the primary containment tank. Several on-line techniques are used to detect, annunciate, and measure releases of fission prod- The most difficult components to mainlain are ucts from failed elements to the primary sodium those under unconventional, relatively inxcessi- and cover gas systems. The fuel element rupture ble, intrinsically radioactive, and radioactively detector (FRED) is used to monitor a small bypass contaminated conditions. Examples of such com- stream of primary coolant for the presence of ponents are control rod drives, the main core grip- delayed neutron emitters in the exit coolant. per, and the subassembly hold down fixture. Such Another device-the germanium-lithium, argon components penetrate the reactor vessel cover and, scanning system (GLASS)-analyzes a flowing as a result, become highly radioactive through neu- cover gas stream by means of gamma pulse height tron activation. The problem of intrinsic r,%dioac- spectrometry with a GeLi semiconductor detector. tivity, sodium contamination, and fission product plateout make direct repair virtually impossible. In The origin of a given fission product release may such a situation-for example, a malfunctioning be identified by means of a xenon-tagging tech- control rod drive, the entire unit is removed and nique developed at EBR-11. The gas plenums of all replaced. When such components are being experimental fuel elements are inoculated with removed, the system must be decayed for a,Dproxi- small quantities of various but unique mixtures of mately 5 d because of the sodium 24. The compo- xenon. Under failure conditions, gaseous fission nent is then pulled into a shielded pipe is handled products and a portion of the xenon tag mixture are by the building crane. released to the cover gas system. An on-line 'device concentrates the xenon component of the cover gas EBR-I1 has been the only operating LMFBR in by cryogenic trapping and analyzes the xenon frac- the world with metallic driver fuel. Several consid- tion with a mass spectrometer. A comparison of the erations affected the initial design of the IEBR-I1 measured isotopic ratios with those of library sam- driver fuel during the early and mid-1950s: relative ples provides positive identification of a source. ease of fabrication, success of prior metallic fuel loadings in EBR-I, excellent heat transfer proper- Various techniques are used to measure impurities of metals, and superior breeding characteristics ties in the sodium and the cover gas system. On-line of metallic fuels. analytical equipment is used to monitor levels of hydrogen, oxygen, and tritium in the primary and As operating experience accumulated, hurnup secondary sodium systems. The primary cover gas limits were incrementally increased from the origi- system is continuously monitored for carbona- nalvalueof 1 at.% establishedin 1961, to 1.12 at.% ceous materials-principally methane, carbon in 1966, to 1.8 at.% in 1969, and to 2.6 at.% in monoxide, and carbon dioxide. Routine samples of 1975. Beyond 3 at. '70,fuel swelling became a prob- primary and secondary sodium are chemically ana- lem. Pressures exerted at the fuel cladding interface

A-1 13 were so large that cladding strain and subsequent siderable interest are the following statistics: of rupture seemed likely. Such limitations were always approximately 10,000 MARK-I1 elements irradi- evident, even in the early days of design. The inevi- ated to the burnup level of 6 at.% and an addi- table penalties were short operating cycles and inef- tional 17,000 irradiated to 8 at.%, none has failed, ficient use of fuel. However, the impact of the that is, suffered cladding rupture. penalties was reduced by the rapid on-site reprocessing of discharged fuel. On certain occasions, The first use of EBR-I1 as an irradiation facility discharged fuel was reprocessed and returned to the began in May 1965 with the insertion of two experi- core within 29 d. mental subassemblies that contained various structural specimens and prototypal fuel rods Despite the successful implementation of short (PuO2-UO2 and U-Pu alloys). Since that time, the turnaround reprocessing techniques, considerable compliment of elements, capsules, and other speci- effort was devoted to the development of fuels that mens may be broken down as follows: mixed U-Pu could be operated to much higher burnup levels. oxide fuels, 3 189; driver fuel elements (surveillance One concept that seemed particularly promising in and run to failure program) 5173; other LMFBR the early days of operation was the permissive fuels (principally mixed U-Pu nitrides and car- swelling principle. This concept was premised on bides), 755; cladding and structural specimens, the contention that if the fuel material was permit- 1343; absorber materials, 219; and miscellaneous, ted to swell unconstrained, a point would be 360. Mixed U-Pu oxide fuels have been irradiated reached at ~30%AV/V when pores in the fuel to a heavy atom burnup of 19.9%; cladding tem- matrix become interconnected. This permits the peratures of greater than 800°C have been release of trapped fission product gases to the fuel achieved; and neutron fluences of 2.1 x element plenum. The fuel, in a weakened condi- neutrons/cm2 have been accumulated in structural tion, tends to deform into the pores rather than materials. Absorber materials have been irradiated straining the cladding. Proof of the principle to 16.5 x 1021 neutron captures/cm3, and EBR-I1 appeared in the form of an improved fuel element driver fuel has been irradiated to heavy atomic design, the MARK-11. The cladding thickness was burnup of 18.9%. increased from 0.23 mm to 0.31 mm to strengthen the cladding; the pin diameter was reduced from Enrico Fermi Atomic Power Plant. The Enrico 3.66 to 3.3 mm; the sodium-filled annulus was Fermi Atomic Power Plant was the first full-scale increased from 0.15 to 0.25 cm; and the gas industrial fast reactor power plant in the United plenum above the fuel was increased to accommo- States. Some features of the FERMI reactor are date the increased release of fission product gases. similar to those of EBR-11. However, EBR-I1 is of The uranium-235 enrichment was increased to lower capacity, 20 MW(e); the FERMI reactor, 67 wt%, and the cladding was changed to annealed although developmental, was designed for an ulti- Type 3 16 stainless steel. Intensive irradiation sur- mate capacity of 65 MW(e), comparable to an veillance studies on test specimens were conducted, average-size conventionally fueled unit of its time. and, as a result, it was shown that the fuel materials swelled rapidly and made contact with the cladding Reactor Features- The arrangement of compo- at approximately 2 at.% burnup. At this point, the nents within the primary reactor tank is illustrated fuel lattice becomes sufficiently porous to permit in Figure A-78. The reactor core and blanket, as relatively free flow of fission product gases to the well as the fuel transfer system, were contained fuel element plenum. Beyond this point, fuel mate- within an irregularly shaped lower section of the rial remains in contact with the inner wall of the reactor vessel. Attention is drawn to the open cladding. But because the fuel is then porous and region between the reactor vessel and the primary “weak,” and because the driving force for diame- tank, which was filled with graphite shielding. The tral swelling is low, the cladding strain rate is low. upper section of the reactor vessel contained a large Consequently, the lower stress on the jacket permits sodium pool above the core and blanket. Access to much higher burnups to be realized. the core was accomplished by a rotating top shield plug arrangement. Currently the burnup limit for the MARK-I1 fuel is 8 at. Yo. A continuing irradiation surveillance The location of other plant equipment is shown program implies that even higher levels are possi- in Figure A-79. Several connected buildings adja- ble; since endurance testing at EBR-11, up to cent to the reactor building contained steam gener- 18.9 at.% has been successfully completed. Of con- ators, a turbo generator, and control equipment. n

A-1 14 G ra not

A-1 15 LEGEND n

1. Steam-generator house 2. Gastight building, 3. Transfer-cask car 4. Primary sodium overflow tank 5. Reactor 6. Primary sodium pump 7. Intermediate heat exchanger 8. Secondary sodium pump 9. Steam generator 10. Secondary sodium dump tank 11. Control room 12. Turbine generator

Figure A-79. FERMI plant layout.

Performance data were generally based on an ini- core Arrangement-The lower reactor vessel, tial power level of 200 MW(t) planned for the first core, and blanket, shown in Figures A-80 and -81, core. The heat removal system had a design capac- were located asymmetrically in the primary tank ity of 430 MW(t) for subsequent core design. for fuel handling purposes and to permit decay heat

A-116 Sweep mechanism Offset handling i drive mechanism, Hold-down drive mechanism Transfer-rotor Plug drive mechanism drive mechanism Deck structure Shot tank -

Rotating-plug container

Sodium level (normal o3eration)

Exit port Sodium level (refueling) 30-in. outlet Offset handling mechanism

Thermal shield Hold-down column Alignment spider Hold-down plate Deposit port 6-in. radial-blanket inlet Exit port 44-in. core inlet Core support plates Radial-blanket support plates Rotor support assembly Subassembly-transfer pot Thermal shield

Transfer-rotor container Radial-blanket inlet plenum Primary shield tank Core inlet plenum Vessel fler:-plate supports

Radial-blanket i

Figure A-80. FERMI reactor vessel.

removal during shutdown by natural circulation. rods. Each of the core subassemblies (Figure A-83) The coolant flow was upward in its natural circula- contained an upper and lower axial blanket section tion system; the large supply of sodium in the pri- in addition to the central fuel bearing core region. mary system (1 36,000 kg) provided the heat capacity needed to reduce transients. An inner radial blanket region, with prcivisions for 34 subassemblies, was provided outs (de the A cylindrical section in the lower reactor vessel, core. These positiods, identical in design to those in 2 meters in diameter and 1.78 m high, contained the core, could be used for additional fuel subas- the core and square blanket subassemblies. The semblies if needed. The orifice design in ea:h type central core region, which was surrounded by the of subassembly, however, controlled the amount of breeder blanket, was in the form of a 78.7-cm- coolant received. diameter cylinder, also 78.7 cm high. Figure A-82 is a reactor cross-section plan view illustrating the The fuel region of each core subassembly con- core and blanket subassembly arrangement. For sisted of 140 fuel pins, each having a nominal 200-MW(t) power level operation, 105 central lat- length of 83.26 cm. Each pin also had an an outer tice positions contained fuel subassemblies. Within diameter of 0.401 cm, consisting of uranium- this region, 10 positions were provided for control 10 wt% molybdenum alloy with an uranium

A-1 17 Support-plate support structure l-

Figure A-81. FERMI meltdown section.

SUBASSEMBLY CLEARANCE

THERMAL BAFFLES

THERMAL SHIELD INSTALLED IN SEGMENTS

-Oumtity 2 8 I os

5a) Pasibls Storage for Core or Inner-rodiol-blanket Subossemblier 24

FOR HOLD-DOWN 195 ASSEMBLY ALIGNMENT GUIDE (31 REQUIRED Thermal Shielding in Form of Steel AS SHOWN 0 TRANSFER Rods Used For Surveillonco Tubes 3 ,/ROTOR CONTAINER Neutron-source Location NOSNO4 I DATUM LINE 'r MECHANISM Surveillance Subosrembly Locolion -I 1 Totol Subwremblier 871 N @ Oscillator-rod Temporory Localion

Figure A-82. FERMI reactor cross section.

A-1 18 140 fuel pins

M Section through core

I rHandling head

73m Strainer - a Yal L C al -m m0 n Q v) .-e I.:-

Fuel pin\4

Match line 'B'

79 c? lMatch line 'A' -. Figure A-83. Isometric view of FERMI fuel assembly.

A-119 enrichment of 25% U-235. Each pin was clad with cross-section, measuring 6.72 cm on each side and n 0.0127 cm of reactor grade zirconium metallurgi- having a nominal wall thickness of 0.244 cm. The cally bonded to the fuel alloy. Zirconium end caps square subassembly shape permitted a plate type served as closures at the top and bottom. The !ewer fuel to be used in a subsequent core, if found desir- end of each pin was fastened to the support struc- able. At the top of the outer tube was a combina- ture by anchor bars inserted through slots in the tion handling head and holddown contact. A lower bottom end caps. The upper ends were’ free to support section, the nozzle consisting of two con- accommodate changes in length resulting from centric round tubes, was attached at the bottom. temperature changes and radiation effects. As The spring, shown in Figure A-83, permitted the shown in Figures A-83 and -84, the core subassem- thermal expansion of the subassembly structure. A bly was enclosed in an outer stainless tube of square strainer was also provided at the inlet of each core

0”-0‘-38” max out of parallel (0.0005 out in 2.693 in.) 1 3 ‘zx7vGauge line

L2.646 in. sq 4

Section A-A * 2.693 in. sq

Spacers shall be concentric with the center line joining dia’s “X9?-$

Transition-square to round

“Z” “2” 2.243 in. dia

32 h-

urT- 0.0003 in. in 1 in. length max out of parallel Section B-B

Figure A-84. FERMI fuel subassembly (external side view). n

A- 120 subassembly. The upper and lower axial blanket tained a pump and intermediate heat exchanger. regions of the core subassemblies contained 16 ura- The pump-a vertical shaft, single-stage centrifu- drs nium rods 1.0 cm in diameter-3 wtol’o molybde- gal pump rated at 745 L/s and driven by a 1,000 hp num alloy, with the uranium depleted to 0.35% motor-circulated sodium to the core through a U-235. A 0.025-cm-thick stainless steel cladding 35.6-cm-diameter pipe. About 13% of the total was used for these rods. Both the inner and outer flow was pumped through a 15.2-cm pipe to the blanket subassemblies used a lattice of 25 rods, radial blanket (Figure A-85). Core coolant flowed each having the same diametral dimensions as the up from the lower inlet plenum of the cole, both axial blanket rods. In the case of the inner blanket through the reactor core itself, and then to the subassemblies, the support structure included a upper reactor vessel, which served as an exit thermal expansion spring arrangement identical to plenum, shown in Figure A-86. The hot sodium that used in the core units (Figure A-84). Outer then flowed by gravity from this pool to the shell radial blanket subassemblies could fill the remaining side of the intermediate heat exchanger through a 500 positions, shown in Figure A-82. These had a 76.2-cm line. The intermediate heat exc nangers simple support structure, as indicated in Figure A-84. were counterflow shell and tube units constructed of Type 304 stainless steel. The primary sodium Control Elements-In this reactor, poison con- stream, by this time at lower temperature, r-turned trol was used rather than fuel or reflector control. to the primary pump. Additional details on the pri- Details on the control rods can be found in Refer- mary cooling system are given in Reference A-1 14. ences A-1 10-1 13. Available excess reactivity was A system of secondary containment for 1 he pri- intentionally kept small. Furthermore, the fuel mary system consisted of a welded, leIk-type management scheme minimized the reactivity enclosure for the reactor vessel, primary system, changes caused by fuel burnup. The eight safety pump tanks, and the intermediate heat exchanger rods and two operating control rods used Boron,l0 shells. This structure, which protects againsi loss of in boron carbide as the absorber. One of the operat- sodium, is described in detail in References A-1 14 ing control rods used for regulating purposes had a and 115. maximum reactivity insertion rate of 0.67 $Is. The other, used for shim control, was designed for 1 4/ Three secondary coolant loops were provided, min insertion. The eight safety rods, spaced uni- each corresponding to a primary loop. Hlzat was formly as shown in Figure A-82, provided a transferred from the intermediate heat exchanger to shutdown reactivity of over $8, and the two operat- the shell side of a steam generator by sodiuin. This ing rods had a total worth of 92P. sodium was circulated by an 840 L/s pump driven Each of the control elements was housed in a by a 350 hp motor. The steam generators were verti- cylindrical guide tube which, in turn, was mounted cal shell and tube exchangers of single-wall tube inside a square tube having the same outside construction arranged for a combination of cross- dimensions as a core assembly. Heat generated by a flow and counterflow. A number of design i’eatures control rod was removed by sodium flowing from provided protection against any leaks within the the core inlet plenum through pressure breakdown unit. The tube to tube sheet joints, for example, orifices inserted in the section between the two sup- were located in an inert gas space above the sodium port planes. Flow was directed up through the con- to reduce thermal shock. trol rod assembly and then through the annulus formed by the control rod assembly and inner Structural and Shielding Components-Th e reac- round guide tube. tor vessel and internal components are shown in Figure A-78. Core and blanket subassemblies were During reactor operation, the safety rods were contained within the lower reactor vessel; the upper held just above the upper axial blanket by a latch on reactor vessel section, at approximately atmos- each safety rod drive unit. Reactivity adjustment pheric pressure, served as a mixing pool for the hot for startup, power level changes, and burnup com- sodium. The upper section contained a core hold- pensation were made by means of the two operating down device and offset fuel handling mechanisms. control rods. These two rods were raised and low- The holddown device maintained radial alignment ered at different rates, as mentioned above. of the core assembly’s upper end and prevented Hear Transport System-In the primary coolant subassembly movement induced by the uplil’t force system, three coolant loops were used. Each con- of the sodium flow.

A-121 Overall length 96-9/16 in. P .-L 71-112 in. blanket rods

aA 73-7/16 in. wrapper tube s- \ '& ,\, /\ y

.A i3 -. Overall length 96-9\16 in. h) 71-112 in. blanket rods 73-7/16 in. wrapper tube

I nl 3

0.039-in. dia spacer on 9-in. pitch

0 ?

71-112 in. Figure A-86. FERMI reactor vessel arrangement.

Spent fuel and blanket subassemblies were plenum, was arranged so that the thickness of mol- removed from the reactor vessel and deposited in a ten material in the region would be about 3. 18 cm. transfer container by the offset handling mecha- A conical flow guide was installed to prevent a nism. Each subassembly in a transfer pot could be buildup of fuel in the center of the meltdown sec- transported by the transfer rotor container to the tion by dispersing the molten fuel as it enteIed the exit port; then the subassemblies were raised verti- plenum. A detailed description of the fuel system is cally into a cask car by the cask car gripper. Spent given in References A-1 13, A- 116- 1 19. The reac- fuel was unloaded from the cask car in the fuel and tor vessel was surrounded by a graphite n1:utron repair building, where the fuel was placed in decay shield located in a nitrogen atmosphere inside the storage prior to shipping operations. Both the primary shield tank (Figure A-80). The shield sys- hold-on device and offset handling mechanism tem also included a 30.5-cm-thick laminated steel were mounted eccentrically on the rotating plug so thermal shield inside the reactor vessel wall. A sec- the holddown device would be swung away when ondary shield consisting of a 76.2-cm thickness of the handling mechanism was swung over the core. concrete surrounded the primary shield tank and prevented neutron activation of the secondary The lower reactor vessel included provisions for sodium system. A 1-1/2-m-thick biological shield safe containment of molten fuel in case of core outside the containment vessel completed the radial meltdown. This section, located in the inlet shielding system.

A- 123 The containment vessel, shown in Figure A-87, dling mechanisms, safety rod drives, and one pri- was a vertical cylinder with a hemispherical top mary system sodium loop. These components, head and semiellipsoidal bottom head. It housed together with sodium purification, were used in the reactor and the primary system. The inside conducting a full-scale mechanical and hydraulic diameter was 22 m, and the overall height was test of the reactor. The tests were isothermal up to 36.6 m, of which 15.5 m was below grade. Design temperatures of 540°C. The overall results of the specifications based on containment of the nonnuclear tests were satisfactory, though modifi- sodium-air reaction provided for the possibility of cations were necessary. Experience with the use of an internal pressure of 0.22 MPa. sodium was excellent. The operation of equipment in sodium and other environments presented no Operating Experience- Starting in July 1959 and major difficulty. It also indicated the high purity of terminating on May 31, 1961, a series of nonnu- large amounts of sodium (up to 181,000 kg) and clear tests were performed at FERMI on compo-, that the system could be maintained without diffi- nents in the reactor portion of the plant. The culty. The test indicated construction materials so-called test facility included the reactor vessel, used in the system were compatible with the primary shield tank, rotating shield plug, fuel han- sodium.

Figure A-87. Vertical cross section of FERMI reactor building. n

A-124 ......

Starting July 1, 1961, the FERMI plant was possibility of erosion. The OHM was modified to incorporated around the test facility and preopera- strengthen the gripper linkage. The modification tional tests began preparatory to nuclear testing. ensured enough strength would be available to per- Testing to 538°C for a period of about 1 week mit application of impact loads and force the grip- checked out the operation of safety rods, fuel han- per open. The stabilizer foot was redesigned to dling mechanisms, and sodium pumps. In general, ensure the OHM would not be locked over a parti- the system performed satisfactorily. Following the ally raised subassembly. The stabilizer assembly 538°C test, tests of graphite directly around the and its attachment to the rotating tube was reactor vessel (inside the insulation) indicated that strengthened to withstand higher lateral loads. The the material failed to withstand the high tempera- possibility of obstructions interfering with future ture test. Tests indicated the borated graphite did operations was avoided by installing a new mecha- not conform to design specifications and would nism called a sweep arm; it was used to check for have to be replaced. Tests on the remainder of mate- obstructions in the upper plenum above the core rial outside the insulation in the primary shield and blanket. About 2,000 man hours were tank indicated this graphite material also needed expended in the vessel by personnel using protective replacement. It was apparent that moisture and suits within the argon atmosphere to accomplish oxygen control was a must in the use of high- results .A-120 temperature graphite. All graphite in the primary shield tank was replaced with high-density, high- In July 1961, a series of hydrostatic iests at temperature reactor grade graphite. Any boron 5.86 MPa on the No. 2 steam generator tubes indi- used was in the form of boron carbide. The rein- cated leaks. Subsequent checks indicated 7 1 tubes stallation of the graphite shield was a primary fac- were cracked through the tube wall. All of them tor in delaying the reactor’s completion. were located opposite one of the two sodium inlets. Testing the tubes indicated that residual stresses, During the test of the offset fuel handling mech- corrosion effects, and elevated temperatures anism (OHM), it became jammed by malfunction induced stress corrosion cracking. The entire tube and misoperation. The OHM became bent when an bundle of No. 2 steam generator was retubed at the inadvertent attempt was made to move it laterally fabricator. All three units were stress relieked and before it was delatched from a partially raised inspected prior to reassembly. dummy subassembly. An innerlock had been disconnected for the purpose of carrying out the test. In December 1962, a sodium water reaction took The OHM was removed with a lightweight remov- place in the No. 1 steam generator, blowing the able container such as not to violate the inert gas rupture disk installed for just such an event. The integrity of the reactor. After disassembly, it was water was dumped manually by an operator to pre- found that the gripper and stabilizer foot were not vent further introduction of water. The No. 1 sec- in their proper positions. After removal of the ondary loop was drained, and the tube bundle was OHM and reduction of the sodium level, observa- removed and cleaned in ~3 m3 of alcohol. The tion of the core and blanket subassembly indicated largest cleaning operation to 1963 to remove a number of displaced subassemblies. Investigation sodium from one component was performelj with- showed the subassembly heads had stuck to the out incident. Examination showed extensive tube holddown plate fingers when the holddown was damage caused by vibration of the tube against the raised. When the plug was rotated with the heads support structures, as well as erosion from the engaged with the fingers, the fingers and other sub- sodium water reaction during the period between a assemblies were bent. Subsequent lowering of the reaction and the rupture disk blowout. Smadl- and holddown resulted in further damage and in full-scale models in water of the tube bundle and its unseating some subassemblies in the support plate, baffle indicated vibration of the tubes at the causing further flow erosion of the support plate sodium inlet had been as high as 0.635 cni. Baf- holds. For repairs, the reactor vessel was drained of fling and lacing of the tubes were carried out to sodium, and personnel in protective suits entered reduce this vibration to a negligible quantit): the reactor. They removed the center support plate and the holddown plate with its fingers, and rede- A program was initiated to modify the check signed the holddown finger sockets. Other modifi- valves in the primary system. The 40.6-cm check cations were made to the control and safety rod valves induced high-pressure surges when lone of guide tubes, The center support plate had stellite the loops was shut down with either one or two of bushings inserted into the holes to prevent further the loops operating. A program was carried out to

A-125 install new check valves-modified to include dash It was decided to interrupt the test program because pots and springs-to reduce the surge pressure to hydrogen levels in the No. 3 steam generator indi- negligible values. cated possible leaks. Inspection showed numerous leaks existed at the welded joints between the tube Each of the more than 3,000 cans in the rotating and tube sheet. In addition, feedwater control for plug containing graphite was vented by 0.16-cm the steam generator was noted to be a problem holes to the argon cover gas above the sodium in the because of fluctuation in the sodium outlet temper- reactor. The primary system tests at 538°C and at atures but with a temperature amplitude of about lower temperatures released material in the graphite 5°C. Relocation of the temperature detectors elimi- binder to the primary system. This release resulted nated most fluctuations, thus averting any poten- in carbon and other contamination of the primary tial single circuit shutdown while on automatic system. Extensive programs of sodium sampling, contro1.A-124 analysis, and filtration were carried out to determine what contaminants were present and how they The ,tests.of greatest interest were the reactor’s affected the system. Tests of the materials in the stability tests. The test program was developed to primary system indicated some degree of carboni- answer whether an operating characteristic could zation but not sufficient to affect operation of the make the reactor unstable and whether the reactor reactor. fuel bowed inward enough to cause a significant positive bowing coefficient. Oscillator tests Three operational tests prior to criticality on through 100 MW(t) showed no evidence of insta- August 23, 1963, were successfully completed; bility or positive bowing coefficient. After the only minor modifications of the control safety rods power run August 5 through 7, onIy a few other and control drives were required. tests at 67 MW(t) were conducted later in the month. Testing was conducted primarily for The low-power nuclear test program for FERMI obtaining plant test data, as opposed to nuclear was originally scheduled for 182 d of reactor time. testing, in such tests as shielding, reactor vessel tem- The program was to verify the soundness of the perature differentials, and core subassembly outlet basic physics design of the fast breeder. Each indi- temperatures.A-125-128 vidual test was completed within the allotted time. However, delays caused by systems external to the Fuel Melting Incident-During a 100,000-kW reactor, extended testing throughout 1964. The sys- power run in August, temperature measurements tems primarily responsible for the delays were the were made of sodium flowing out of the top of sev- cask car and the steam generators. Twenty-seven eral of the 103 core subassemblies. This was accom- tests were scheduled at low power; all but two were plished by reading the output of thermocouples completed by the end of 1964. The reactor was installed in the holddown finger. The thermocou- taken critical over 250 times, using 500 different ples were installed just above the subassembly han- core configurations that required movement of dling heads in about one-fifth of the core and inner 1,400 subassemblies and more than 3,500 move- radial blanket subassemblies. Such measurements ments of the fuel handling mechanism and rotating allowed increased sodium heat to be determined as shield plug.A-122 the sodium flowed through and reduced the heat in a given subassembly. This increase in heat content Throughout the tests, the plant operated very could then be checked against the calculated heat stabIy and was easiIy controlled. Predictions of the generation rate of a given subassembly for consist- reactor’s parameters were, in general, substanti- ency. Such comparisons were not taken to be pre- ated. The predicted isothermal coefficient was cise. High precision would also require an accurate reduced by 7%. Reactivity worths of the control determination of the sodium flow rate through rods were found to be 8% lower than predicted, and individual subassemblies-something that was not safety rod worths were 15% lower than directly measurable. predicted .*-l 23 Abnormal temperatures first appeared in June, The test plan for the nuclear program during when the outlet sodium temperature for one subas- high power ascension comprised the oscillator, sembly was as much as 19°C higher than expected. pulse, power coefficient, automatic reactor pro- The temperature returned to normal during the gramming, and scram tests. Results of all tests July run and then ran high again during the August through 20 MW(t) showed satisfactory operation. runs. While unexplained phenomena are always of n

A- 126 concern, the higher-than-expected temperature ting down the reactor. Experience at other facilities readings were in all cases lower than the normal indicated that several of the previous reactclr mal- sodium outlet temperature of the hot test subas- functions involved anomalous negative rea'ztivity. sembly located at the center of the core. Therefore, Erratic behavior of the dn/dt meter that had been observed temperatures were all well within designed experienced before was only 1 or 2 MW/min, and margins of the subassemblies and 480°C below the was thought to be a noise pickup in the contr,ol sys- boiling point of sodium. It was decided to continue tem. Indeed, it soon disappeared. During the inci- the test program but to shift the misbehaving subas- dent, the reactor was placed on manual control, but semblies to different locations. It could then be was returned to automatic following a clearup of determined whether the temperature anomalies the dn/dt signal .A-129 were the result of a fault in the fuel subassembly itself, a fault in the thermocouple, or a fault in the At 3:05 p.m., the feedwater flow control 'system associated environment. was put on automatic. Variations on the dn/dt meter were observed. It was noted that the two As part of the continuing plant test program, it operating control rods appeared to be withdrawn was planned on October 5 to bring the reactor to an further than normal for the approximately indicated thermal power level of 74,000 kW, a level 30 MW(t) power level that the reactor had now surpassed several times before. Several tests were to attained. This indicates an unexplained source of be performed, including checking a new pressure negative reactivity; one possible cause would be control adjustment on the main steam bypass line, higher-than-expected heating. So, the power adjusting the automatic feedwater control, and increase was halted, and a check was made of sub- measuring temperatures on the troublesome No. 1 assembly outlet temperatures. During this :heck, steam generator, across the core outlet, and near alarms sounded from the radiation monitors in the the transfer rotor. In addition, oscillator tests were upper reactor building ventilation exhaust ducts, planned because operation was to involve a particu- resulting in automatic isolation of the containment lar combination of two loops that had not been building. The area radiation monitor in the Fission tested at this power level. Information was to be product detector building also exceeded its set- obtained on the previously observed unexpected point, isolating the fission product detector system. subassembly outlet temperatures by observing the A Class-I radiation emergency was announced. temperature of the three deviant subassemblies in Reactor power reduction was begun in accordance their new locations under thermocouples that had with operating procedures, and by 3:2CI p.m. previously indicated normal temperature readings. assessment of the reactor information was complete and the reactor was shut down by a manual The reactor was started up the night before and scram, with insertion of all six safety rods into the kept at a 1.0 MW(t) over night. It was intended to core, terminating the chain reactions within begin a power ascent at 8 a.m. on October 5, but a se~onds.~-l29 malfunction in a steam generator valve postponed beginning the power increase until 1:45 p.m. At The following information was developed during 2:15 p.m., with the power at 5 MW(t), difficulty the first few days after the accident, which demon- was experienced in the control system of the east strated that the extent of fuel damage was low. boiler feedwater pump, and the power was reduced to 2 MW(t) to start another pump. At 2:20 p.m., Freedom of movement of the six safety the rise in power was began and continued until rods was verified the next day. The overall 8 MW(t) was reached, when there was a brief hold structural integrity of the core was not to put the reactor on automatic control. An auto- damaged enough to cause gross destruc- matic system took the reactor up in power at a con- tion of the safety rod guide tubes. trolled rate and involved a somewhat ' complex withdraw pattern of the two operating control rods. Primary coolant system flow was maintained without primary coolant leakage, During the automatic startup, the meter record- indicating no credible mechanism for mas- ing the time and rate of change of power (dn/dt) sive core melting. gave erratic signals. A similar but independent device was connected to the safety system so that if A subcritical reactivity check performed the measured power decrease exceeded 6 MW/s the on October 7 indicated there musi: have safety rods would drop into the core, thereby shut- been some net fuel motion away from the

A-127 center of the core, and that the movement were in their correct position to within the tolerance n was equivalent to the complete removal of of the OHM and the subassemblies. Then, by out- about half the fuel of an average fitting the OHM with a dial indicating lighter sub,assembly. spring gauges, it was used to establish the pulling force required to lift each subassembly a specified Review of the chart of the core subassem- distance. bly outlet temperature recorded suggested that overheating and corresponding fuel Very high lifting forces were required in the area melting were located in or near one core of the two subassemblies shown by thermocouple assembly (M- 140). readings to be running hot during the October 5 run. It was determined that the two subassemblies 0 On October 20, samples of the core cover were stuck together. gas were extracted and sent to Argonne National Laboratory for analysis. Based .$ 'After the decay heat had diminished to a point on the analysis, it was concluded that the where the sodium was no longer needed, the amount of fuel melted was almost equiva- sodium was drained from the reactor vessel by a lent to that contained within one complete siphon procedure to a level 2.5 cm below the under- suba~sembly.~-~~~ side of the top core support plate. The core was viewed through a boroscope inserted through one Recovery operations must include removal of the of the access ports of the rotating plug at the top of subassemblies, and normal subassembly removal the vessel. A sharp wedge on an extension handle requires raising the holddown mechanism. This was lowered between the subassemblies to chisel the device is mounted on top of the rotating plug and subassemblies apart. On July 1967, the bond extends all the way to the top of the core, for indi- between the subassemblies that were stuck together vidual fingers exert a downward force against each was broken and the subassemblies were removed core subassembly, holding them in place against the and shipped to a hot cell for examination. upward force of the flowing coolant (Figure A-80). Such operation might have caused core movement. Surface examination of these two subassemblies revealed definite damage. They were bent and there Two more investigations were conducted to shed were holes in the wrapper can. Though the weights further light on the extent of core damage and to had not changed, the center of gravity had shifted assess the potential for secondary criticality acci- downward, 1.75 cm in one subassembly and 3 cm dents before subassembly removal procedures were in the other. begun. One investigation involved a pressure drop measurement, which was carried out on One subassembly was cut longitudinally; the October 21, 1966; it was concluded that flow other was cut transversely in four places. A signifi- blockage in the core ranged between zero and six cant portion of the fuel was found to have melted subassemblies completely plugged. and sloughed downward, though a number of fuel pins remained intact. No evidence could be found In November of 1966, shortly after the pressure to indicate that molten fuel had flowed from one drop measurements were completed and chances subassembly to another. for a secondary criticality situation were assessed as negligible, the holddown mechanism (HDM) was In August 1967, the sodium was drained further raised. A special procedure was used incorporating to below the level of the primary meltdown section, strain gauges and acoustic and reactivity readings which effectively emptied the inlet plenum of to detect any irregularities as the HDM was raised sodium. In September, a boroscope viewing device in small increments. and light were lowered through access ports in the rotating plug down to the meltdown pan located The mechanical arm of the sweep mechanism 10.7 m below the operating floor. was operated manually and passed over the top of the core and then over the blanket subassemblies. On the zirconium liner of the meltdown section No protruding subassemblies were indicated. was an object appearing as a crumpled, folded piece of sheetmetal. Stereoscopically viewed, a An offset handling mechanism (OHM) was used three-dimensional map of the object was devel- to confirm that all subassembly handling heads oped. And from the map, a model was made of n

A-128 - . - . - ......

what the object probably looked like before crum- communication with the other operator was by pling. The unfolded version appeared to be flat, phone. The object was pulled by the retrieval tool pie-shaped piece of sheetmetal about 1 ft long, up 10.7 m of the 35.6-cm line at a rate of 1.5 m which was identified as one of six zirconium liners every 20 min to allow cooling of the segment so it of the conical flow guide. It had become detached would not burn the plastic glove box. Examination and was thrust up against the inlet flow nozzle of confirmed its tentative identification as a zirco- the effected subassemblies by the force of the flow- nium segment, originally installed in the conical ing liquid sodium coolant. flow guide.

Repair of FERMI I- Removal operations were The remaining five zirconium segments were conducted in two phases. The first phase involved detached by removing the three tack welded screw removal of the detached zirconium segment. The heads holding each plate and melting them with an second involved both detachment and removal of electric arc. Following detachment, the segments the remaining five segments still believed to be were then retrieved by the same method used for the bolted to the conical flow guide. first segment. An arc melt tool was designed that fit through the same 4. l-cm holes in the support plate, Plans for the first phase were initiated immedi- that could be remotely operated in a hot argon ately on discovery of the object, even before it was atmosphere, and that could be sufficiently articu- identified. The approach taken was to retrieve the lated to allow positioning over each of the screw object intact through the 35.6-cm primary line, heads. Such a tool was successfully tested on the aided by a manipulator inserted through the rotat- full-scale mockup in August 1968. ing plug and support plate holes. The spine tool manipulator, resembled a human spine. It con- In November 1968, the arc melting tool was low- sisted of a 91.4-cm length of a series of round disks ered into the inlet plenum and the screw removal with egg-shaped cross sections. A series of cables process began. The proper gap between the elec- allowed bending in any desired direction and con- trode and the arc melting tool and the screw to be trolled a jaw at the far end that was used for grasp- melted was determined optically by obserking the ing the object. Control was at the operating floor length of a high-frequency low-current spark dis- level, some 12.2 m away from the jaw end. charged from the electrode. When the propx position was obtained, a 1350-amp dc arc was initiated The retrieval device was a 1.5-m section of 12.7-cm- for 14 seconds. All of the screw heads were thus diameter flexible metal hose attached to a 12.2-m melted without incident before any of the segments length of helical spring cable. A remotely controlled were retrieved. Following screw head removal, the grabber was mounted on a trolley within the flexible segments remained in position, slightly welded to hose. The retrieval device was inserted into the 35.6-cm the flow guide by the arc melting operation It was primary sodium line through a 15.2 x 17.8-cm hole necessary to insert a special chisel between the seg- cut into the pipe just outside the secondary shield ments in the conical flow guide and to slightly twist wall, 10.7 m from the object to be removed. A spe- them to pry them free. Removal of the segments cially designed glove box was attached at the position from the vessel was then accomplished without of the opening in the pipe to maintain the inert atmo- incident by an retrieval device improved from the sphere of the primary system. Figure A-88 shows the earlier retrieval experience. On December 16, 1968, relative location of these devices. about three weeks after the first screw was removed, the last of the six zirconium segments was The tools and procedures were all checked in a retrieved. full-scale mockup of the lower plenum and the 35.6-cm inlet sodium line. Following retrieval of the segments, the arc melting. tool was used to weld the zirconium screw In March 1968, the first phase of retrieval began. shanks to the conical flow guide. The inlet plenum Tool operators were positioned at the top of the region was inspected for debris and cleaned by a rotating plug and at the glove box over the opening vacuum. The primary system was again filkd with in the 35.6-cm sodium pipe. These operators were sodium, followed by cold trapping operations that separated and were not in visual contact (see caused impurities to precipitate in the cold trap, Figure A-88). Only the spine tool operator atop the thus maintaining acceptable sodium purity. In a rotating plug could view the lower plenum region; very real sense, damage to the reactor caused by the

A-129 Figure A-88. Perspective view of the segment removal operation. n

A-130 October 5, 1966, fuel melting incident was now bly was loaded into the core, these inserts, called repaired. Before resuming operations, the follow- flow guards, extended approximately 2 inches ing was done: below the support plate. These flow guards were basically a hollow cross with a pointed tiottom. A photographic examination of the metal- Design studies showed that the possibility of an lic fingers of the holddown mechanism object blocking flow through these inserts was nil. assured that distortion of the damaged subassemblies had not damaged the hold- The steam generators presented a problem dur- down mechanism ing the preincident test: gross instability whenever the temperature of the sodium entering the steam All core subassemblies in the reactor dur- generator reached the saturation temperature of the ing the melting incident were replaced by water. This instability caused severe thermal cycling new subassemblies stored on site at the water header and twice caused the plant to shutdown. The problem was solved by inserting Added to the instrument system to aid in smaller tubes inside the original downcomei- tubes. the operation and detection of steady state With water flowing inside the smaller tubes, heat reactor anomalies were a digital computer transfer resistance from sodium was increas1:d such (malfunction detection analyzer), two that boiling could not start until the water entered delayed neutron detectors that could detect the risers. The steam generators showed no indica- a high power level failure area equal to 3% tion of instability in all the tests after modification. of 1 fuel pin, a very accurate ammeter connected to the reactor power monitors, and Once the cause of the fuel melting incident was digital readouts for the thermocouple determined, and modifications were instadled to accurate to 0.06"C. ensure that such an incident could not happen again, planning began for resumption OF plant The accident demonstrated that thoughtful prep- operation. aration for malfunctions, coupled with a design philosophy that incorporates redundancy and in- One of the first items to be determined was which depth defense, provides a safeguards envelope that fuel subassemblies should be used. There were three will adequately respond to a real accident and pro- types of fuel for the ongoing program: (a) fresh, vides an ample margin between actual conse- unused fuel, (b) fuel with recorded proof that it quences and public harm, even though reactor was not overheated, and (c) fuel that was exfamined accidents seldom progress exactly as envisioned by metallurgically analyzing a tip cut from the end during design. The maximum credible accident cap on a fuel can and carefully measuring the sub- defined for FERMI I is complete blockage of one assembly for distortion (called requalified fuel). subassembly at full power; the AEC staff defined it These three types of fuel were of sufficient quantity as blockage of two subassemblies. The predicted, to operate the reactor for a substantial period. most likely cailse of the accident, is rapid melting and injection of fuel from the core, followed by It was decided that after completing a power automatic scram from the resulting rapid power escalation program, the plant would be operated to drop. Consequences of such an accident poses no do the following: hazard to public or plant personnel. The real accident involved a slower, partial melting of only part Demonstrate plant operability of the fuel in two adjacent subassemblies. It was detected by anomalous reactivity loss and radiation Further debug and tune the equipment release to the containment followed by manual scram. The consequences posed no public or plant Obtain revenues from the sale of electricity personnel hazard. The containment system was designed to accommodate a secondary criticality Develop and demonstrate the tec:hnique accident involving collapse of one half of the total for routine fuel reloading and removal core into the other half. The real accident resulted from the sodium-filled system. in melting less than 1% of the core. To ensure operational readiness, all equipment Inserts were designed for the bottom flow nozzle underwent a formal, complete calibration and of each subassembly. When a modified subassem- operational check before plant operation was

A-131 resumed. More than 50 preoperational tests were Last minute design changes must be sub- written, reviewed, conducted, reported, and ject to the same design quality and review approved in this program. By the end of June 1970, as the original design. The basic cause of all preoperational testing was completed; 93 core the FERMI m,elFdown was the detachment subassemblies had been loaded into the core; seven of one of' the' zirconium segments added central core positions were temporarily occupied by after construction was well underway. Had blanket subassemblies; and 12 core subassemblies there been more careful design review and were in storage ready for exchanging with those in model testing of this late addition, tech- the core lattice, all to provide reactivity necessary niques could have been used to preclude for criticality in the 200-MW(t) operation and dem- their detachment-or they might not have onstration program. been added at all.

The power demonstration program began on Despite environmental extremes within the November 20, 1970, with a 5-1/2-d continuous core of a sodium-cooled reactor following operation period. During the run, 8 160 000 kWh operation, perseverance and ingenuity and of electricity were generated. Following the run, the special tool design can permit extensive in- plant operated on an intermittent basis for contin- reactor repairs. ued testing of equipment. By June 28, 1971, the decrease in core reactivity owing to burnup had Considerable thought should be given in reached the point where the maximum power designing nuclear plants to provide access attainable with all rods withdrawn was 163 MW(t). for major repairs, should they be necessary. At this time, the maximum burnup in the central core subassembly was 0.28 at.%. Reactor accidents usually are preceded by anomalous, though sometimes complex, It was decided that it was too early to send the sub- reactor behavior. If a plant instrumentation assembly for the examination scheduled for 0.4 at.% system is designed to detect and help inter- burnup. Rather, changes first were made to the core pret such behavior, the probability of occur- configuration to increase its reactivity, so that rence of such accidents can be reduced. For 0.4 at.% burnup could be attained while operating at example, the anomalous temperatures and full power. Up to this point, the plant had generated reactivity preceding the FERMI meltdown 23 370 000 kWh of electricity. Changes included led to subsequent installation of a malfunc- adding several new subassemblies and removing the tion detector analyzer. oscillator equipment. Because some evidence of subassembly bowing was observed following the 1966 The ability to preserve a continuing record fuel melting incident, the decision was made to mea- of plant data is a valuable tool for diagno- sure the degree of bowing in every subassembly prior sis and assessment of plant system to lowering the holddown mechanism. response in the event of an accident. Proc- Additional information drawn from the melting ess computers offer immense capability incident follows: for such a task.

Distributed and redundant safety rod Local geometry and off-normal operating capability is basic to ensuring shutdown conditions should be carefully considered during an accident. The scram following in designing thermocouple installations the meltdown involved insertion of six and interpreting their output. safety rods; one rod would have been sufficient to shut the reactor down. However, During the incident, the maximum temperature c there could well have been rod binding in reading of 379°C was below the normal full-power the safety rod channel near the meltdown bulk outlet sodium temperature of 428"C, and far region. Whereas even partial insertion of below the sodium.boiling point of 882°C. One sub- the rod during the FERMI incident could assembly melted, yet a neighboring subassembly probably have shut the reactor down, it is did not. Heat conduction to neighboring subas- reasonable to assume that subassembly semblies, rather than convection through the low- melting could produce sufficient distor- flow partially blocked subassembly, became the tion to preciude insertion of a nearby rod. predominant heat removal mechanism. n

A-132 The outlet sodium temperature was further Idaho National Engineering Laboratory (INEL) reduced by conduction of the surrounding sodium near Idaho Falls, Idaho. Radioactive primary as it flowed through the hole in the subassembly sodium remains at the site awaiting delivery to the handling head to the thermocouple; thus, the tem- Clinch River Breeder Reactor project. Noni-adioac- perature rise through a partially blocked subassem- tive secondary sodium was sold and shipped to bly as measured by an outlet thermocouple is no West Virginia and processed into sodium methyl- longer inversely proportional to the flow rate for ate; the primary cold and hot traps, including mis- low flows. As the flow to such a subassembly is cellaneous piping, were shipped to Beatty, Nevada, decreased, the measured coolant temperature will for burial; the least enriched uranium was sent to reach a peak and then drop. Thus, whereas depar- the Savannah River Plant; the blanket subassem- tures from normally observed temperatures at a blies were sent to the subsurface storage area of the given power are indicative of flow abnormalities, Radioactive Waste Management Com plex at such temperature readings in FERMI could not INEL; other miscellaneous radioactive or contami- directly indicate a condition of sodium boiling or nated items were buried. The last category includes fuel melting. dummy subassemblies, oscillator rods, neutron sources, special reactor and pool handling tools, Renewed Operation at FERMI-During the recov- etc. Whenever possible, miscellaneous hardware, ery phase, one of the basic decisions was to modify such as pool racks, were decontaminated and sold the plant so as to be absolutely certain that a condi- as scrap. When this was not possible, they were tion resulting in fuel melting could not occur again. shipped for burial. The problem was attacked by removing the possibility of a flow blockage, and equipping the control The plant boundary was revised to exclude many room with instrumentation to detect such a block- nonradioactive areas, such as the office and turbine age at a very early stage, before melting could take buildings. The new boundary is marked by a 7-ft- place. After completing the plant changes, the high chain link fence, and building walls that power demonstration program was resumed. enclose the fuel and repair building, reactor building, health physics building, sodium storage build- The measurements were completed on ing, and cold trap room. The cold trap room August 26, 1971; the holddown mechanism was contains some contaminated piping. The sodium lowered on August 3 1; reactor operations resumed storage building contains three 57-m3 tanks of pri- on September 1. The period culminated with a 10-d mary sodium. The health physics building will high-power run, from November 20 to remain functional and within the enclosed area December 1, 1971. Total power generated by the until the primary sodium has been shipped off site. plant from critically in 1970 to this point was There are also 630 barrels of primary sodium tem- 32 260 000 kWh. At completion of the run, the porarily stored in the reactor building. The reactor maximum burnup in the core reached 0.4 at.%. vessel contains a nitrogen and carbon dioxide cover gas. This atmosphere extends to caps on secondary After its best operating run to date, the FERMI sodium pipes external to the reactor building, to reactor shut down December 1, 1971, for a 2- to piping within the reactor building, and into the pri- 3-month period for destructive examination of a mary shield tank. Cover gas pressure is maintained fuel subassembly and application for a license to go by a bottled gas system. The overflow tank was iso- from 0.4 to 0.6 at.% burnup, a probable core lated on the surface system, passivated with C02 limit. The reactor supplied 9 million kWh to the and opened to the air atmosphere of the lower reac- Michigan power grid from November 19 through tor building. The sodium service system piping was December 1, 1971. Of the 142 h of operation in cut in the cold trap room and in the reactor bidding that period, 80% was at the full 200 MW(t) and caps were welded onto the pipes where they licensed power level. connect to the primary system. Sodium service piping between the reactor building and the co Id trap Decommissioning of FERMI-The major retire- room was closed by valves at one end and by welded ment activities and plan for maintaining the plant caps at the other end, wherever they were cut. The in perpetuity formed the framework of the decom- insides of the pipes are Contaminated with a very missioning program. The buildings, reactor vessel, thin layer of sodium. One of the tunnel lines con- major primary sodium piping, and many auxiliary tained sodium and this line was sealed with a facilities remained intact. Blanket elements were welded cap in the lower reactor building. The auxil- sent to the Idaho Chemical Processing Plant at the iary fuel storage facility was sealed after

A-133 100 pounds of C02 were added to passivate any respond to water levels by the insertion of copper n residual sodium that may have drifted from the fuel probes 3 inches into the floor drain. A similar mod- storage pots. ification was made to the moisture detector in the fuel and repair building hot sump. The moisture Because all irradialed structural materials and all monitor in the lower reactor building remains large quantities of sodium that cannot be drained intact. are behind shielding, there were no radiation levels in the above-floor region of the reactor building A carbon dioxide cover gas is maintained in most above 0.1 mR/h, except from the barrels of systems containing residual radioactive sodium. sodium, and none in the below-floor region above Cover gas pressure is maintained above atmos- 0.3 mR/h. The fuel and repair building transfer pheric pressure and below 35 KPa. The cover gas tank and overflow tank were drained, sealed, and system is checked during inspection of the facility. passivated with C02, then opened to the air atmo- A backup supply of gas is available to supply the sphere of the tank room. If the building is disman- system for eight days under normal conditions. tled, these tanks will require special handling since Extensive details on the deactivation and surveil- they still contain a small yield of residual sodium. lance of FERMI I can be found in Reference A-131. The four circulation cold trap systems associated with the transfer tank have been completely It is a notable achievement that no worker was removed, disassembled, and shipped for burial. ever exposed to an excessive effluent release during the life of the FERMI project. This is in the face of The three 57-m3 tanks in the sodium storage build- the extensive maintenance and repair work required ing contain a total of 144 m3 of sodium. The maxi- in an extremely difficult environment following the mum radiation level in December 1973 was 9 mR/h 1966 fuel melting incident. at the installation surface. Once the sodium was FFTE The Fast Flux Test Facility (FFTF) is a transferred to drums, the tanks were passivated 400 MW(t) LMFBR designed and constructed for with C02. At some future date, the tanks may be irradiation testing of breeder reactor fuels and cut up, scraped, or shipped for burial, depending materials. The reactor provides extensive capability on the difficulty of decontamination. The cold trap for in-core irradiation testing, including eight core room is vacant except for sections of service piping. positions that may be used with contract instru- Radiation levels in the cold trap room in mentation attached to all or part of the test speci- December 1973 were 2 mR/h, and contamination men. Four of these positions may be used for was less than 100 d/m/decimeter2. Access to the independently cooled (closed) loops. tunnel between the reactor building and the cold trap room can only be obtained by removing a In addition to irradiation testing capabilities, the welded cover. From the cold trap room, the tunnel FFTF provides long-term testing and evaluation of is closed off by a concrete barrier. Radiation levels plant components and systems for liquid metal fast in the tunnel have not been measured, but contami- breeder reactors. The fast test reactor is equipped nation levels in December 1973 were less than with structures for heat removal, containment, core 100 d/m/decimeter2. The liquid waste in sump component handling and examination, instrumen- pump system has been deactivated but left intact so tation and control, and for supplying utilities and any potential ground water leakage can be pumped other essential services. The reactor is located in a from the sump to the fuel and repair building liquid shielded cell in the center of the containment. Heat waste storage tanks for later discharge when a suf- is removed from the reactor and liquid sodium cir- ficient quantity has accumulated. culated by three primary loops including primary pumps, piping, and intermediate heat exchangers Access to the FERMI facility is through locked (also located in containment). The secondary gates and fencing within the exclusion area of the sodium loop transports the reactor heat from the present nuclear and fossil power plant. All gates are intermediate heat exchangers to the air cooled tubes locked except when personnel are inside, and a vis- “dump” heat exchangers (DHX). Heat from the ual inspection of the fence and building is made two closed loops is removed by similar but much weekly to determine that all gates and doors are smaller heat transport systems. properly closed. The FTFF has capabilities for receipt, condition- The moisture detector in the biological shield ing, and storage of a reactor core, and core compo- wall cavity in the reactor building was modified to nent and test assemblies are routinely installed and n

A- 134 ...... -. .. . .

removed. Unlimited examination and packaging are fuels. These positions are cooled by the main reac- also available (for outside shipment). For a detailed tor coolant. For each of these positions, coolant description of the FFTF facility, see Reference A-132. output flow and temperature are measured by instrumentation suspended above the position. Reactor Features-The reactor vessel is constructed of Type 304 stainless steel. It is approxi- Nonfissioning materials may be tested in the mately 13.1 m high and has a 6.1-m inside fueled portion of the core and also in some of the diameter. Its wall thickness varies from 6 to 7 cm. A 108 core positions surrounding the fueled zone. pool of liquid sodium fills the reactor vessel from The neutron flux in these peripheral positions vary the bottom to the apron cover gas zone near the from 0.5 to 4 x IOl5 n/cm2/s. Nominally, the top. Sodium coolant enters the reactor vessel 108 positions contain Inconel reflectors. through three 40.6-cm nozzles in the lower part of the vessel, flows through the core and other equip- The core periphery positions may also irclude a ment, and then out directly through three 71.1-cm variable number (up to 15) absorber assemblies. outlet nozzles located approximately at the vessel They remain fixed in the core periphery for reactiv- mid-height (Figure A-89). ity adjustment during a particular fuel cycle. The other peripheral positions are occupied by Inconel The reactor vessel is suspended from its upper reflectors, which are surrounded by radial shielding section and nested in a bottom-mounted Type contained in a core barrel. The core barrel supports 304 stainless steel reactor guard vessel whose pur- six core restraint mechanisms that hold the core pose is to ensure that the reactor vessel nozzle assemblies in proper orientation during nuclear remains submerged in sodium in the event of a leak. operation. The core barrel is supported by a core The reactor head is the closure for the reactor vessel support structure welded to the reactor vesel. The and is 7.62 m in diameter, about 56 cm thick, and reactor internals consist of the reactor core, three weighs 195 tonnes. Equipment mounted on the instrument trees that support flow and temperature head includes drive mechanisms for in-vessel com- measuring instruments over the core, nine nuclear ponents, seals, and shielding for openings, thus control drive mechanisms, radial shielding around providing access to the vessel interior and coolant the core, the core support structure, a core barrel piping for independently cooled test loops. that contains the core, six core restraint mechanisms, a horizontal baffle and seal to (control core Arrangements-The FFTF reactor core sodium flow in the reactor vessel, three low level generates up to 7 x 1015 n/cm2s of neutron flux flux monitor thimbles, and three in-vessel storage (60 to 65% greater than 0.1 MeV) at a power level modules. The locations of these components are of 400 MW(t). shown in Figure A-90.

The core comprises 199 core assemblies. The The reactor core and other reactor interrials are design includes 74 positions for PuO2-UO2 driver immersed in a pool of liquid sodium. The liquid fuel assemblies, which generate the nuclear flux; sodium fills the free volume inside the reactor vessel nine boron carbide control rod absorber assem- from the bottom to a cover gas zone near the top of blies; and eight positions that may be used for inde- the vessel. The sodium coolant enters the reactor pendently instrumented tests of fuel specimens or vessel through three 40.6-cm inlet nozzles, flows reactor core materials. All eight independently through the core and other equipment, and out instrumented positions may be used for open tests through three 71.1-cm outlet nozzles. cooled by the reactor primary coolant system. Four of these positions may be used for independently Each driver fuel assembly, which consists of cooled closed loops. In these closed loops the cool- 217 individual fuel pins, can be used for statistical ant system is completely separated from the FFTF testing of fuel pins under actual operating conditions. heat transport system. This permits testing of fuels Each assembly can also be used for operational test- and materials over a wide range of temperatures in ing of alternative designs and materials for the nonfis- a controlled environment. sioning components of fuel assemblies.

In addition to the testing capabilities of the open The core will also accommodate up to eighst inde- test and closed loops, the 74 positions designed for pendently instrumented test assemblies. They may driver fuel may be used for statistical testing of be used for testing either fissioning maierials,

A-135 structure

Figure A-89. FFTF reactor cutaway.

A-136 Figure A-90. FFTF reactor internals

A-137 advanced fuel, or nonfissioning materials. The access to the core for refueling, the instrument trees independently instrumented test assemblies may must be rotated away from the core into a stored include up to four independently cooled closed position. Before such rotation, the control rod 0 .. loop in-reactor assemblies; The independently dr:ve lines must'be disconnected remotely at the top instrumented test assemblies are 12.2 m long. The and bottom of the instrument tree guide tubes. The other core assemblies are 3.7 m long, except for 1 actual nuclear control absorber assemblies remain some slightly shorter reflectors. in the core.

The three instrument trees (see Figure A-91) are a The core support structure, which is welded to unique feature of FFTF since they cover the core; the reactor vessel (Figure A-93), positions the core the control rod drives (Figure A-92) pass through components within the reactor vessel. The core bas- guide tubes in the instrument trees. To obtain ket is centrally located within the core support

vo 4

rotational drives

Removable instrument assembly guide tubes

-Top of core

Figure A-91. FFTF instrument tree schematic.

A-138 Control rod drive mechanism

tl- Shielding 11

Upper disconnect

Instrument tree upper frame

Instrument tree

Lower disconnect Botton of instrument tree guide tube

Control rod absorber assembly - Top of core ducts

Figure A-92. FFTF nuclear control system.

A-139 Reactor vessel /wall ._

Core -support structure

structure and provides vertical support for the The core restraint mechanisms are mounted on the driver fuel, test, absorber, and inner reflector core barrel. Six mechanisms are equally spaced assemblies. The basket is a closed cylinder with around the periphery of the core and are actuated 151 tubular receptacles connecting the upper and by shafts extending through the reactor head. lower tube sheets. The horizontal baffle mounted at the upper end Different sizes and shapes of receptacles are used of the core barrel acts as a thermal insulator and to ensure that the core assemblies are properly flow barrier between the hot outlet plenum of the placed. The cylindrical section of the basket has reactor vessel and the relatively cool inlet plenum. 12 rectangular flow slots fitted with flow strainers The baffle is sealed to the reactor vessel liner by the through which the incoming sodium flows. The baffle liner interface seal. core support structure and core basket are made of Type 304 stainless steel. In-vessel storage modules are located in three sections of the annular region between the core bar- The core barrel is a Type 304 stainless steel cylin- rel and the reactor vessel thermal liner. Each mod- der about 0.35 m high and 3.5 m in diameter. The ule (Figure A-95) has 19 natural convection-cooled core barrel surrounds the radial shield of the core Type 304 stainless steel receptacles for core compo- and is supported by and welded to the core support nents and one transfer port position for the core structure. component top.

The radial shield consists of six inner blocks and six Assemblies from material surveillance samples outer shield blocks arranged around the periphery of are also located in the in-vessel storage modules. the radial reflector region. The inner and outer shield During reactor operations, these samples are blocks are composed of vertical Standing-Type 304 exposed to the sodium, thermal, and radiation stainless steel plates. The radial shielding is supported environment. Figure A-96 shows the location of by the core support structure. two of the in-vessel storage modules.

The core restraint mechanisms (Figure A-94) The low-level flux monitoring (LLFM) instru- apply radial force to the load pad of the core assem- mentation for determining the reactor’s reactivity blies to hold them in position for power operation. status during subcritical and low-power conditions n

A- 140 . . - .. .. . -. . .. . -. . .. -.

Horizontal baffle standoff

I inner module Yoke shank Outer module

inner shield plates

Core barrel \ Lower Static yoke ring

Figure A-94. Arrangement of core strain mechanisms around core.

comprises three assemblies. The assemblies are open test assemblies, positions for independently located 120 degrees apart in the reactor vessel. instrumented and independently cooled closed loop Each assembly (Figure A-97)is made up of a com- assemblies, boron carbide neutron absorblm for position thimble, a sensor drive unit, a sensor with nuclear control, and Inconel@ connectors. cable, and a nitrogen gas cooling system that maintains the neutron sensor in its cabling below 150°C. Whereas the general design of all 199 core assemblies is similar, details of the handling sockets The drive mechanism, mounted above the vessel and nozzles differ to ensure proper location of vari- head, positions the neutron sensor axially within ous types of assemblies in the core. the thimble. The sensor is located at the core mid- plane during reactor shutdown and startup opera- The core is arranged in concentric hexagonal tions. Prior to full power operation, the sensor is rings of assemblies surrounding a central assembly. retracted to a higher elevation. Rows one through four are known as the inner enrichment zone, rows 5 and 6 the outer enrich- The FFTF core contains PuO2-UO2 driver fuel ment zone. These first six rows constitute the core’s assemblies, positions for independently instrumented fueled zone, which is 0.91 m in axial leng,th and

A-I41 Horizontal n baffle

Core support structure

planum

Figure A-95. Location of in-vessel storage modules.

Thermal liner

Transfer port position Module shell

Storage tube

Figure A-96. In-vessel storage module.

A- 142 Pressure housing - & Connectors Wire harness - assembly %!.,Retainer Dlua

Gear motor 'coup1 i ng Seals Seals Outer shield 'plug assembly ! r 7p4- Outer thimble

Support thimble Baffle thimble

Bearing Core mid-plane -'Sensor support subassembly

Figure A-97. Low-level llux monitor assembly.

1.2 m in equivalent diameter. These six rows con- of the three heat transport system cooling circuits tain the driver fuel assemblies, nine control rod (see Table A-10). assemblies, and up to eight independently instrumented test assemblies. The neutron flux control system operate:; three primary control rods and six secondary c,ontrol control System-The reactor plant control sys- rods. During normal operation, the three primary tem provides equipment for stable and reliable con- rods are fully withdrawn while the secondary rods trol, both manual and automatic, of the reactor are used for control. The primary and secondary and heat transport systems during normal opera- rods have separate control equipment. The flux tion. Routine controls include removal of decay controller provides stable, automatic control of heat during reactor shutdown and neutron flux reactor power, which is accomplished by operating control, and also coolant flow and the dump heat the secondary control rods using flux as thc feed- exchanger bulk sodium outlet temperature for each back signal. The power setpoint is adjusted by the

A-143 Table A-10. Reactor plant control systems n

Controlled Function Variable Sensor Actuator

Flux control Neutron flux Compensated ion Control rod chamber

Primary flow control Primary loop sodium Magnetic flow meter Liquid rheostat of flow primary pump motor

Secondary flow Secondary loop Magnetic flow meter Liquid rheostat of control sodium flow secondary pump motor

Secondary cold leg DHX bulk sodium Thermocouple Air flow (fan speed temperature control outlet temperature and/or damper position)

operator. The feedback signal is the auctioneered six-coil six-phase step type that requires dc to ener- high of the three secondary power range nuclear gize the coils. signals. The controller also provides secondary rod withdrawal stops based on high flux and high flux- The sodium flow control system includes all flow ratio. As a result, all secondary rods are equipment necessary to provide stable control of blocked for withdrawal when the allowable flux or sodium flow in each heat transport system loop. flux-flow ratio is exceeded. The control is obtained during all modes of operation by varying the pump speed through use of liq- The control rod selection logic ensures that only uid rheostats. The system includes a flow setpoint one control rod in the primary system or one in the generator and six control subloops-one for each secondary system can be manually withdrawn at a pump. Feedback signals are derived from the loop time It also ensures that the automatic control can flow instrumentation. operate only one secondary rod at a time. Rod posi- The flow controller output drives the liquid rheo- tion is measured and indicated by two completely stat position control system. The liquid rheostat independent methods: a relative broad position provides the capability to vary the motor resistance indicator and an absolute rod position indicator. of a wound rotor motor, thereby changing motor speed. A speed ratio of approximately 2:3:1 is The major elements of the control rod system are available, providing flow control over a range of the flux controller, the control rod selection logic, about 50 to 100% of design flow. the rod position indicators, the motor generator (MG) set, the control rod drive mechanism The dump heat exchanger (DHX) control system (CRDM) controllers, and the power supplies. provides stable control of airflow, thereby maintaining limits of the DHX sodium outlet tempera- Separate MG sets, circuit breakers, and 3.6-phase ture. Included are interlocks and equipment transformers are provided for the primary control rod required to rapidly reduce airflow after reactor and secondary control rod systems. scram.

The MG set provides three-phase power for input Each of the 12 DXH modules has a control sub- to the three- to six-phase transformers, which in loop to operate the fan’s variable speed control and turn provide the power sources for the CRDM con- the controllable outlet dampers (a coarse control trollers. These controllers supply pulsed dc to ener- damper with six blades and fine control damper gize the CRDM stator coils. The CRDM motor is a with two blades). n

A- 144 The reactor plant control system provides an temperature of 360"C, inlet pressure of about integrated design to monitor, control, and coordi- 0.917 MPa, and a nominal outlet temperature of nate the decay heat removal process for plant shut- 530°C. The actual inlet and outlet temperatures down operations. Following a scram or shutdown depend on the operating conditions selected. to refueling temperature, the decay heat of the reactor and stored heat in the HTS are dissipated by the primary and secondary loops at pony motor flow Dry argon cover gas is used to blanket the conditions. The DHX modules (Figure A-98) are sodium in the reactor vessel and throughout the cooled by natural air circulation. heat transport system to avoid contact between sodium and air. The heat transport system has three The reactor plant control system also provides 133-MW stainless steel circuits (Figure A-99). Each the monitoring control and coordination functions circuit consists of both a primary (radioactive) and necessary to ensure decay heat removal under com- a secondary (nonradioactive) loop. The reacior ves- pletely natural circulation conditions if all electri- sel and its head are part of each primary loop; addi- cal power (off-site power and on-site emergency tional primary loop equipment consists of a power) were lost. Special instrumentation, separate primary pump with two motors, two isolation from the normal control system, has operated from valves, one check valve, required piping, and a shell a Class IE battery power source. Manual control of side of the intermediate heat exchanger (IHX). The the fine damper is provided from the control room IHX isolates radioactive primary sodium from the for each of the 12 modules; the DHX outlet tem- secondary loop. perature for each module is displayed adjacent to the manual controller. Each main component and each vertical run of piping in the primary loops are provided with a Heat Transport System- Heat generated by the guard vessel. The primary loops transport reactor 400-MW(t) reactor is removed via the heat trans- heat to the intermediate heat exchangers that ther- port system (HTS). The HTS pumps 27 400 L/s of mally link the primary and secondary loops. The sodium through the reactor vessel at a nominal inlet three primary loops have common flow paths

odium inlet

Variable damper

Sodium outlet

Preheater Fan solation doors Variable speed dr cs Figure A-98. Dump heat exchanger module. A-145 I Sodiumoutlet I -P

c , c, Inlet 2 plenum

Figure A-99. Sodium flow through FFTF reactor vessel.

through the reactor vessel but operate indepen- Heated sodium flows from the reactor outlet nozzle dently otherwise, except for a cover gas equaliza- through the hot leg isolation valve to the section of a tion line common to all three loops, linking all centrifugal free surface pump (Figure A-100). primary pumps with the reactor vessel. Sodium from the pump discharge is circulated to the shell side of the intermediate heat exchanger where Each secondary loop consists of the tube side of the heat is transferred to secondary sodium. From the the IHX, a secondary sodium pump with two intermediate heat exchanger outlet, the primary motors, an expansion tank, and a sodium-to-air sodium flows through a check valve and a cold leg dump heat exchanger (DHX). The DHX consists of isolation valve. It then goes to the reactor vessel inlet four 33-MW heat exchanger modules per loop, and and through the reactor, thus completing the flow required piping and isolation valving. cycle. Covered guard vessels around each primary pump and IHX retain the sodium in the likely event of The sodium coolant enters the reactor vessel a leak. below the bottom of the core, flows up through the core and reactor internals, and out slightly above Piping from the reactor vessel to the pump sec- midplane of the reactor vessel (Figure A-99). tion is 7 1.1 cm OD; the remaining piping from the

1 A-146 Cold leg piping (16 in.)

* Provides atmospheric seal between HTS cell and HTS pipeway

Figure A-100. Heat transfer system primary loop piping. pump discharge to the reactor vessel inlet nozzle is The pump is driven by a 2,400-V ac-wound rotor 40.6 cm OD. The decision to make the reactor ves- induction motor. The motor is hand-controlled by sel outlet pipe 71.1 cm was governed mainly by the a liquid rheostat variable speed control. An auxil- frictional pressure losses in this line. iary 480-V ac pony motor on each primary pump performs low-flow decay heat removal at low power The sodium flow rate is measured by a and standby conditions. The pony motors are permanent-magnet flowmeter located downstream driven by the FFTF emergency generators if normal of the IHX outlet; flow is adjustable by a variable power fails. speed drive pump. Maximum fluid velocity is less than 9.14 m/s, to limit coolant cavitation, pipe Typical primary coolant pump design features erosion, and pipe vibration. The primary hot leg are as follows: piping (Le., reactor to IHX) is Type 316H stainless steel. The cold leg piping (Le., IHX to reactor) and components are Type 304H stainless steel. A design flow of 915 L/s of sodjum at 1.24 MPa and 566°C Equipment and piping containing primary sodium are installed inside steel-lined shielded cells that contain inert gas (nitrogen) during normal A variable speed ratio of 2.3: 1, which pro- reactor operations. vides flow control over a range of about 50 to 100% of design flow Each primary pump (Figure A-101) is a vertical, single-stage, single-suction, controlled inlet, free A stable head capacity curve between surface centrifugal pump located in the primary flows of 26.5 m3/min and 68.1 m3/min at hot leg piping. The suction nozzle is 78.1 cm in design speed, with drawdown of l.!i m or diameter, the discharge nozzle 40.6 cm in diameter. less at design flow

A- 147 Q

Oil seal-bearing

Figure A-101. Heat transfer system primary pump with guard vessel.

A-148 A seal system to prevent escape of radioac- Design pressure of 1.65 MPa on the pri- tive argon from the primary system, or in- mary side and 1.82 MPa on the seEondary flow of lubricants or fluids to the pump side Maximum pressure of 70 kPa on the pri- Optimized coastdown characteristics to mary side and 100 kPa on the se1:ondary minimize thermal transients and to pro- side at design flow rate vide adequate reactor core cooling after a scram Effective tube length geometric center of at least 4.6 m above the reactor core gc:ometric Pony motor drive to provide about 7.5% center, ensuring adequate thermal driving of design flow at 9.9 kPa with automatic head for natural convection cooling takeover from main motor to pony motor during coast down, preventing any flow The three secondary loops, shown in Figure A-103, interruption circulate nonradioactive sodium coolant to transport heat from the tube side of the IHX to the air-cooled Completely drainable system. dump heat exchangers (DHX). No direct sodium interconnection exists among the loops. Each HTS pump has a lubrication circulation system to carry away heat generated by frictional Heated secondary sodium leaving the IHX flows losses in the seals and bearings. to the air-cooled DHX through 40.6-cm-diameter piping via a containment penetration. Near the The IHX (Figure A-102) is a vertical counterflow DHX, the piping branches and reduces to :!0.3 cm design. It transfers reactor heat from the primary diameter and connects to a 20.3-cm-diameter hot loop to the secondary loop. This is achieved via the leg DHX isolation valve. From this valve, the pipe primary flow outside the tubes and secondary flow connects to an inlet header of an individual DHX inside the tubes. module (of which there are four to each secondary loop, 12 total for the heat transport system). Primary sodium enters the IHX below the top of the upper tube sheet and is directed around the tube Cooled sodium from the outlet of each module bundles down along the outside of the tubes. A flows through a 20-cm-diameter cold leg DHX iso- small percentage of flow (about 1%) is continu- lation valve; the piping then expands from 20 to ously returned to the primary pump tank through a 40 cm. Finally, it connects to the 71.1-cm-diameter 5.1-cm orificed line. As a result of this return, gas suction line of the secondary pump. From the collection in the upper regions of the primary pump discharge, the sodium piping enters the HTS (shell) side of the IHX is prevented. The primary cell through a penetration in the reactor contain- 0 sodium leaves the bundle area just above the bot- ment building and connects to the IHX. At that tom tube sheet. It then flows into the annulus point, the flow path is completed. Sodium flow rate between the shell and bottom plenum and out the is measured by a permanent-magnetic flowmeter nozzle in the shell hemisphere. located downstream of the pump discharge. The flow rate is adjustable by a variable speed pump Secondary sodium enters the top of the unit, drive. A Venturi flowmeter, located upstream of the flows down through the central downcomer into secondary pump, is used for flow calibration of a the lower hemispherical plenum, and up through secondary-loop permanent-magnetic flowmeter. the inside of the tubes into the upper plenum area. The secondary flow then leaves the IHX through The 40.6-cm-diameter secondary hot leg piping the horizontal outlet nozzle. and the section of 71.1 cm cold leg piping at the inlet to the secondary pump are of Typc: 316H One IHX is located in each of the three HTS cir- stainless steel. The 40.6-cm cold leg and the 20-cm cuits downstream of the primary pump and ahead piping and components are of Type 304H stainless of the check valve. steel.

To meet the system operating requirements, typi- The secondary piping loops are designed for a cal design features in the IHX are as follows: pressure of 1.72 MPa throughout. This design

A- 149 n

Figure A-102. HTS intermediate heat exchanger and guard vessel.

A-150 ......

Secondary pump

16 in. secondary

IHX shell side-primary tube side-secondary

4 dump heat exchanger (DHX) modules for each loop

Figure A-103. HTS secondary loop schematic.

pressure allows the secondary loop to operate at a 100% of design flow and accommodates a sodium high pressure in the IHX, so that if a leak occurs in level change of 0.91 m caused by seconda1,y loop the IHX tubing, the leakage would be from nonra- sodium volume change. The pony motor drive pro- dioactive sodium of the secondary loop into the vides 10% of design flow at 0.91-m head, with primary loop. automatic takeover from main to pony motlx during coastdown to give uninterrupted flow. An expansion tank having a surge volume filled with argon gas is provided in each secondary loop DHXs are sodium-to-air exchangers that transfer to accommodate sodium thermal expansion and to heat from the secondary loop sodium to the ambi- pressurize the system. Secondary loop sodium present air (Figure A-105). One DHX unit of four mod- sure at the intermediate heat exchangers is always ules is located in each of the three secondary loops. maintained above the primary loop pressure. Each DHX module is equipped with a fan having variable inlet vans, fan-drive motor, fine and coarse Each secondary sodium cooling pump is a verti- control dampers, isolation gates, manual and auto- cal, single-stage, single-suction, controlled inlet, matic controls, tube-bundle, oil-fired preheaters, free surface centrifugal pump (Figure A-104). Its ducting, stack, and airflow turning vents. The design is similar to that of the primary pump, nominal thermal rating of the DHX module is except that the secondary pump has a shorter, 33 MW at 32°C inlet air. smaller shaft and a slightly smaller impellor. Simi- lar main motors and pony motors are used. The The DHX tube-bundle is made up of a biink of second pump is located in the secondary cold leg 66 four-pass serpentive tubes located between hori- piping between the DHX and the IHX. zontal upper and lower headers. The four-pass ser- pentine tubes consist of thinner sections 9.1 rn long The secondary sodium pump is designed for joined by bare U-2 return vents. Suitable baffling at 915 L/s sodium flow at 440°C in 1 MPa. It is the side and ends minimizes bypass of air. Turning designed for a variable speed ratio of 2.3:l which vans are provided to ensure uniform airflow over provides flow control over a range of about 50 to the tube-bundles.

A-151 Figure A-104, HTS secondary pump.

A-152 ...... -...... - .. .- . .- .. - - . w .:::::st I .:::::st I - 'DHX preheater air flow path .f- DHX normal operating air flow path

urn blower air flow Oil return

Figure A-105. HTS DHX Air flow diagram.

The air supply system has a double-width Structural and Shielding Components double-inlet centrifugal fan with airflow control provided by a variable speed coupling (to vary van Containment-The reactor containment sys- speed) and by variable inlet guide vans. The fan is tem consists of a cylindrical carbon steel reactor driven by a 1,250-hp electric motor It provides air- containment vessel 56.9 m high by 41.2 m in flow of 280 kg/s with air side pressure drops of diameter. Also included are several principa 1 struc- 2.75 kPa of water. Other design features incorpo- tures and equipment pieces within the vessel. Steel- rated into the DHX to meet the system operating lined reinforced concrete cells occupy the lower requirements include the following: portion of the containment vessel from grade level to approximately 78 ft below grade. DHX tube-bundle geometric center 9.1 m above the IHX geometric center to ensure In the operating area, a shielded operating floor adequate thermal driving head for natural is located at grade level. convection cooling, if required Cells and Pipeways-Below the operating floor are the cells and pipeways (Figures A-106 Sufficient air draft through the and A-107). These internal structures are made of tube to dissipate a Of reinforced concrete. Cells and pipeways thai. house 10% of design power piping and/or equipment containing primary system sodium are provided with an inert nitrogen gas Instrumentation to detect any sodium atmosphere. Steel liners contain the inert atmos- leakage. phere and protect the concrete from expo,jure to sodium in case of a spill. Cells containin,: large Each DHX bundle is preheated through a circu- quantities of sodium are designed with liners lation system with air heated by an oil-fired burner. backed with insulating brick. These heaters establish or maintain a temperature of 200°C during standby when sufficient decay or The reactor cavity, located in the center of the pump heat is not available. containment vessel, houses the reactor vesyel, the

A-153 .'

Primary Na vessel cell

Containme

Figure A- 106. Typica! plan of internal structures within the containment vessel.

Containment y vessel

Figure A-107. Typical cross section of internal structures within the containment vessel.

A-154 . .- .. . . - .. - .. . - .. .

guard vessel, the main reactor support structure, locks are provided so that neither a door nor its associated shielding, and reactor pi$ng. The cavity equalizing valve can be opened unless the other is insulated and temperature controlled to maintain door is locked and sealed. Also, a welded construc- allowable temperatures for concrete. Three HTS tion hard patch on the containment vessel can be cells are adjacent to the reactor cavity on the east, cut open if needed. south, and west sides. These three cells house the primary sodium pump and intermediate heat The equipment air lock is used for all transfers of exchangers as well as associated equipment, piping fuel into and out of containment. It will also be and instrumentation for each of the three primary used for transfer of large equipment if it is needed heat transport system loops. for maintenance.

Four closed loop systems (CLS) cells and associ- The equipment air lock is approximately 7.6 m ated pipeways are located between the east and in diameter with doors at opposite ends, each with south HTS cells and between the west and south a clear opening 3.4 m wide by 5.1 m high. An HTS cells. The CLS cells are designed to contain emergency escape hatch in the air lock provides an pumps, heat exchangers, piping, equipment, and opening 61 cm in diameter. This air lock (extends instrumentation associated with the CLS primary into the reactor service building. removal systems. The personnel air lock is 3.05 m in diameter and The northwest quadrant of the containment, 3.66 m long, with doors 2.03 m high and 1.07 m below the operating floor, is the location for the wide. This air lock extends into the Auxiliary interim decay storage (IDS) vessel cell. Also found Equipment Building East in an area convenient to in this area are cells for the auxiliary liquid metal the control room. system equipment, piping, and valves. The emergency air lock is 1.83 m in diameter and Cell access closure consists of shielding access 3.66 m long doors 0.62 m in diameter. This air lock plugs, rolling shield plugs and steel shielding doors, extends into the HTS Service Building West, below and soft shielding plugs or shielding blocks, as grade level. applicable. Operating History-FFTF initiated its first oper- Shielding access plugs are steel-lined (sides and ating cycle April 16, 1982. Full power, steady state bottom), tapered, and stepped concrete plugs that operation began April 23 following a planned 36 h fit into an embedded steel frame in cell or pipeway hold at 92% power to allow restructuring of new ceilings. These plugs are equipped with seals to fuel in the core. On May 14, a small fission gas leak contain the inert cell atmosphere. was detected and subsequent analysis identified the faulty element. After 30 d of continuous opera- Rolling shield plugs are steel lined, tapered, and tion, the plant was automatically shut down due to stepped concrete plugs mounted on wheels that roll an inadvertent valve operation during routine on tracks in the floor. Cells utilizing rolling plugs maintenance. During the shutdown period alter the have bulkhead type, gas tight doors behind the scram and during a routine start of primary sodium plugs to contain the inert cell atmosphere. pump, P-1,a flashover occurred in the bru!;h/slip ring of the pump motor. The resulting pump over- Soft shielding plugs are made of concrete cast speed and flow imbalance among the three primary into a steel plug frame in a cell wall. The soft plug is coolant loops forced the sodium into the shaft removed by breaking up the concrete mechanically annulus. This contributed to the shaft binding on and is replaced by casting new concrete in place. the primary sodium pump, P-3.The upper portion The cell liner is continuous and must be cut out to of the pump was disassembled, revealing a reduced secure cell entry. Soft plugs provide for future shaft clearance in the pump baffle area. The access to areas that have no present requirements reduced clearance was attributed to sodium residue for maintenance access. in this area. The residue came from an unplanned transient caused by a high sodium level in the pump An equipment air lock and two personnel air tank. The pump was freed by heating the shield locks are provided for access. Each lock has two plug and pump tank. After reassembly, vibration doors in series. Each door is equipped with redun- and coast down measurements were taken to verify dant seals and a pressure equalization valve. Inter- no additional problems existed.

A-155 On August 26, the FFTF returned to power oper- supplementing the standard tag isotopic ratios with ation to complete the first operational cycle, fission product ratios and by using a novel probabi- Cycle 1. During the startup some DHX fan speed listic treatment, a tentative location of the leaking Q control problems occurred resulting in an isolation test pin was made. . of a DHX module. A shutdown was conducted to return the DHX module to service. After returning Fission gases continued to be released from the to power, a reactor scram occurred during a routine leaking pin. But amounts were too small to make a calibration of the primary system flowmeters. On positive tag identification until May 21 at approxi- September 12, the reactor obtained 100% power mately 1500 h, when another large spike of fission and remained at full power until November. In gas was released. This gas release occurred about November, a special shutdown evolution was con- 1 h after reactor power was reduced to 98% ducted to investigate the phenomena of the reac- because of high ambient temperature. This time, tor’s pressure drop. On November 11, 1982, after tag gas,concentrations were large enough to make 53 consecutive days at full power, the FFTF’s first an absolute determination of the tag; tag T-12 was operating cycle was completed with a planned reac- confirmed in two successive cover gas samples. tor shutdown. Upon making an identification, FELT drafted a statement of results and recommendations to During the early part of Cycle 1, the pressure present to the plant manager. In addition to identi- drop around the primary system began to increase fying the test assembly as containing the leaking slowly. This increase continued throughout the pin, FELT recommended some supplementary sur- operating portion of the cycle. An engineering and veillance activities to be performed while operating operations effort was expended evaluating and test- with a known leaker in a core. ing this anomaly. No cause has yet been identified, and investigation will continue into Cycle 2.A-133 Plant operations continued until 1320 h on May 24, when the pump controller failure scrammed On May 13, the cover gas monitoring system the plant. Concurrent with the scram, an extremely detected a small burst of fission gas released to the large burst of fission gas (roughly 25 times greater cover gas volume. This initial signal was not large xenon-I33 activity than the May 21 burst) was enough to trip the fission gas high alarm; a cover observed. Analysis of cover gas samples taken at this gas sample taken later in the day was found not to time showed only tag T-12 was present; this confirmed contain any detectable fission gases. The next day, the AB-1 test assembly contained the leaker. During at approximately 1800 h, a much larger burst of the shutdown, the AB-1 test, which contained the fission gas was observed; this event was large leaking fuel pin, was replaced by a driver fuel assem- enough to trip the fission gas high alarm, and bly. (Reference A-134 describes the event in detail). follow-up action was taken. Plant operators took three cover gas samples, and the failed element During Cycle I a total of 9.69 rn3 of solid radio- location team (FELT) was notified. FELT was com- active waste with a total activity of 19.8 mCi was posed of members from physics and irradiation transferred from FFTF for processing. The waste testing, chemistry and analysis, nuclear analysis, originated from the following activities: core evaluation, and FFTF plant operations. The FELT was responsible for coordinating leaker- Interim examination and maintenance cell locating activities. Another of its responsibilities decontamination and maintenance was making recommendations to the FFTF plant manager regarding identification of the leaking Maintenance and storage facility decon- subassembly and follow-up actions. Cover gas sam- tamination and maintenance ples taken on May 14 failed to give a definitive indication of tag gas or fission gas from the leaking pin. Sodium removal system maintenance

On May 15, a third and still larger burst of fis- Heat transport system south sampling and sion gas was observed beginning at about 300 h. maintenance This burst was approximately three times larger than the May 14 event. A cover gas sample was Miscellaneous maintenance in the plant taken and analyzed; the amount of detectable tag in this sample was only one-tenth of the amount con- Radioactive liquid waste sampling and sidered necessary to make a tag identification. By load out station.

A-156 Also transported were 8 m3 of liquid waste with Radiation surveys conducted in support of radia- a total activity of 1.76 nCi. This liquid was gener- tion work in the plant showed direct levels ranging ated from maintenance activities on the sodium from background to about 1 rem/h. Contan9iination removal station in the reactor service building. levels ranged from background to about 60,OCD dpm/ dm2. The 1 rem/h direct radiation was associated General area radiation levels during Cycle 1 were with primary sodium cover gas sampling evolutions. very low with dose rates measured in routinely The radiation level was at contact with the gas tag occupied areas less than 0.2 mrem/h. The few sample trap prior to installation of the shipping cover. exceptions to the design levels involved low level, With the shipping cover in place, the contact radia- small area streaming through shield penetrations. tion level was 80 mrem/h, and personnel exposure For example, the maximum shield contact rating during the sampling evolution was limited to about was 8 mrem/h on a shield located about 3 m above 5 mrem. The 60,000 dpm/dm2 contamination level lower level. Other streaming locations were equally was associated with fuel handling and was a direct inaccessible or in areas with extremely low reading on a sodium contaminated floor valve. occupancy. Removable contamination on the floor valve was 5,000 dpm/dm2. Routine radiation survey data are summarized in Table A-1 1. The values presented were extracted An entry into Cell 489 on July 8, 1982 showed from weekly routine survey reports. These values general area radiation levels to be less than do not reflect special situations such as entry into 0.5 mrem/h. A maximum reading of 1 rnrem/h cells containing primary sodium and maintenance was obtained in contact with a pipe containing pri- on radioactive systems. Special radiation surveys mary sodium. were performed for these operations, and the data are discussed below. Routine radiation contamina- An entry into Cell 490 on June 26, 1982 showed tion surveys were also conducted. No contamina- general area levels to be less than 1 mrem/’h. The tion was detected in the routinely occupied areas. maximum loose contamination level in Cell 490 Local contamination control areas were established was about 5,000 dpm/dm2. Radiation arid con- for special tasks being performed under radiation tamination levels associated with the interim exami- work procedures. nation and maintenance cell, the sodium removal

Table A-11. Summary of routine radiation survey data during FFTF Cycle 1

Dose Rate (mrem/h)

Dose Rate General Area Maximum Maximum Dependent Average Local During Local On On Reactor Location . During Cycle Cycle 10/30/82 Power Comments

I. RSB, 500, Radioactive 0.5 4 0.5 No 4 mrem/h due to temporary storage of IEM Cell components.

2. RSB, 550, General Area 0.2 1 1 No 1 mrem/h due to floor valve adapter.

3. RSB, 520, Rm 204 0.2 0.2 0.2 No Looking for potential leakage to secondary chilled water from IEM Cell Removal System.

4. RSB, 520, General Area 0.2 0.2 0.2 No Local dose could be dependent on cladding breach and release of fission gasses.

5. RSB, 543, General Area 0.2 0.2 0.2 No Same comment as for 4.

6. RSB, 533, General Area 0.2 4 0.2 No 4 mrem/h due to small leak of 41Ar from RAPS instrument line.

A- 157 Table A-11. (continued) n

Dose Rate (mrem/h)

Dose Rate General Area Maximum Maximum Dependent Average Local During Local On On Reactor Location During Cycle Cycle 10/30/82 Power Comments

7. HTS-S, 550, General Area 0.2 0.2 0.2 NO -

8. HTS-S, Rooms 481, 493, & 482 0.2 2 2 Yes 2 mrem/h due to valve stem penetration (P-633) into Cell 489, reactor at 100% Source 24Na.

9. HTS-S, 550, Room 485 0.2 3 2 Yes 2 and 3 mrem/h due to 24Na radiation streaming around shield plus to Cell 492-E, reactor at 100%.

10. RCB, 580, Mezzanine 0.2 3 1 Yes 1 and 3 mrem/h due to contact measurement on low level flux monitor cooling system. Source is transported activation products and 16N.

11. RCB, 550, General Area 0.2 0.5 0.2 No General area dose rates in RCB at full power are not significantly different from background. 0.5 mrem/h in Primary Sodium Pump 1 Pit.

12. RCB, Head Compartment 0.2 6 6 No 6 mrem/h due to depleted uranium shielding in two small local areas. General area 0.2 mrem/h.

13. RCB, 520, Room 540 0.2 0.2 0.2 NO -

14. RCB, Stairway 517 (West) 0.2 0.6 0.6 Yes -

15. RCB, 500, Room 564 0.2 0.5 0.5 Yes 0.5 mrem/h at locked barrier to restricted access radiation zone.

16. RCB, 520, Room 554B 0.2 0.2 0.2 NO -

17. RCB, 500, Room 562 0.2 0.2 0.2 NO -

18. RCB, 500, Room 570 0.2 9 9 Yes 9 mrem/h due to 24Na radiation streaming around shield plug to Cell 567 (EM pump cell), reactor at 100% power.

19. RCB, 520, Room 553B 0.2 0.2 0.2 NO -

20. RCB, 500, Room 565 0.2 0.2 0.2 NO -

21. RCB, 520, Room 5368 0.2 8 6 Yes 8 mrem/h due to 24Na radiation streaming through penetration in a small, localized area about IO feet above floor.

22. RCB, 500, Rooms 561 & 563 0.2 0.5 0.5 Yes 0.5 mrem/h due to 24Na radiation streaming around shield door to 532C.

23. RCB, 520, Room 5528 0.2 0.5 0.5 Yes 0.5 mrem/h at locked gate to 552A. source is 24Na radiation streaming into 552A from HTS-1 (521).

A-158 system, and the radioactive liquid waste system secondary sodium pumps and the air blast heat were low-less than 18 mrem/h and less than about exchangers (see Figure A-108). 10,000 dprn/dm2 removable contamination. JOYO is fueled with a mixed oxide of plutonium Plant personnel radiation exposures were low, as and uranium. The reactor is a loop type and has shown in Table A-12. The values shown are within two identical cooling circuits, each having a heat the statistical variation in the dosimetry system for removal capacity of 50 MW(t). Each circuit con- very little doses and are at preoperational levels. sists of a primary loop, intermediatl? heat The maximum individual dose for a quarter was exchanger, and secondary loop. The heat generated 220 mrem and was associated with maintenance in the core is finally dissipated to the atmosphere personnel working at another reactor facility. The through sodium to the airblast heat exchangers. highest exposure for personnel not associated with Reactor power is controlled by two regulation rods work at another reactor facility was less than and four safety rods. 50 mrem for a quarter. Reactor Features-The reactor vessel is :I stain- Japan less steel vessel with an inner diameter of 3 6 m, a wall thickness of 25 mm, and a height of approxi- JOYO. The experimental fast breeder reactor, mately 10 m (Figure A-109). The vessel is tlouble- JOYO, is Japan’s first experimental LMFBR. It is walled, and nitrogen gas fills the space between the designed to serve as an irradiation test facility and two walls. The gas serves to warm the vessel when to provide construction and operational experience introducing liquid sodium into the vessel, as well as for future LMFBRs. JOYO has been constructed preventing fire by accidental sodium leakage on the site of the Oarai Engineering Center of through the inner wall. The lower part of the vessel PMC, located on the Pacific coast approximately has two inlet nozzles for sodium coolant. The 193 kilometers north of Tokyo. upper part has two sodium outlet nozzles and some other smaller nozzles, such as those for an auxiliary The JOYO plant consists of four main buildings: cooling system, argon cover gas system, iind an reactor building, main cooler building, waste disposal overflow system to maintain the sodium level in the building, and maintenance building. The reactor is vessel constant. A spent fuel storage rack is located below ground in a containment vessel. In installed near the reactor core within the vessel. addition to the reactor itself, the reactor building con- Outside the storage rack, there are steel irradiation tains the principal components of the primary cooling shield plates to attenuate the radiation flux to the system, such as the intermediate heat exchanger and reactor vessel. Just inside the vessel, there is the primary sodium pumps. The main coolant build- another rack on which to place the irradiated speci- ing contains the secondary cooling system with the mens of the vessel material for surveillance lesting.

Table A-12. FFTF personnel exposure summary during Cycle 1 operations

1st Quarter 2nd Quarter 3rd Quarter CY-82 CY-82 CY-82

Number in mrem/ Number in mrem/ Number in rnrem/

Group Average Group Person Average Group Person Average Group --Person

Operations 105 27 99 9 105 4

Maintenance 143 16 138 16 134 2

IEM Cell 22 4 23 4 33 1

Others 65 16 61 7 97 6

A-159 n

IF GL* 200 c

0 2F GL- I%O(

GL-ZO.O(

A. 6. C. PRI. dump tank I D. PRI. overflow tank E. Fuel handling machine F. Large rotating plug L. Main control room G. Control rod drives M. AUX. air blast cooler H. Fuel chargeldischarge machine N. Main air blast cooler I. Cold trap 0. Blower J. EM flowmeter P. SEC. dump tank K. Transfer rotor Q. SEC. main pump

Figure A-108. JOY0 cutaway view.

Thermocouples are located above the reactor core vessel, there is a graphite shield approximately 1 m for measuring coolant temperatures at each core thick. A steel safety vessel surrounds the graphite fuel subassembly outlet. A double rotating plug shield and the reactor vessel and will contain the serves as the lid of the reactor vessel. The function reactor core under sodium coolant during any pos- of the plug is to shield radiation emanating from sible accident condition. the core and also to position various devices such as the refueling machine. The large plug has a diame- Core Arrangement-The reactor core consists of ter of approximately 4.7 m that can be rotated core fuels, blanket fuels, control rods, and a neu- & 180 degrees. The small plug has a diameter of tron source all made up of subassemblies 2.9 cm 2.9 m and rotates from 0 to 180 degrees. Both long of identical hexagonal cross sections. Integral plugs have freeze metal seals to keep gas tightness type core fuel subassemblies are employed; the during reactor operation. The plugs are designed length of the core fuel zone is 60 cm, and the length for pressure and shock equivalent to an explosion of the upper and lower blanket fuel zone are both of 50 kilograms of TNT. Total thickness of the 40 cm. The core fuel subassembly contains 91 fuel plugs are 2.5 m and the upper part of each plug is pins, 6.3 mm OD, 1.91 m long, separated from cooled by nitrogen gas. each other with a wrapping wire. The blanket subassembly contains 19 fuel pins, 15 mm OD, and an All control rod drive mechanisms are mounted effective link of 1.4 m. The fuel pins contain fuel on the small plug and extend down to the core pellets of mixed plutonium and uranium dioxide, region for driving the neutron absorbers. Refueling and the blanket pins contain fuel pellets of depleted and inspecting inside the vessel are accomplished uranium. The control rod subassembly (two regula- through holes in the small plug. Outside the reactor tion rods and four safety rods) contains a neutron

A- 160 Fuel charge-discharge

Control rod drive

plug driving unit Large I'otating plug . . driving unit

,I Coolant (outlet

.I .- ...... - Reactor vessel

Safety vessel .. C. . .._ .. .. *. Core support ..

Coolant inlet -- .* ,. . .* - ' -.. . . :*. . _...... '. .. - '*. :. ..1- 7...'... Figure A-109. Vertical cross section of reactor and core configuration of JOYO.

A-161 absorber of boron carbide enriched 91 070 with facility. It is initiated by the detection of a high neu- boron.1° The neutron source subassembly is of tron flux level, short reactor power oscillations, low antimony-beryllium and located in the radial blan- coolant flow in the primary loops, loss of power, an n ket region. For proper neutron economy and shield- earthquake, or whenever the reactor containment ing, reflectors are arranged in the outermost region isolation mode is automatically tripped. The con- of the core (Figure A-110). tainment isolation mode is initiated by the detection of high temperatures, high pressure, or high Control System- JOYO's core is so designed radiation levels in the reactor containment vessel. that the reactivity coefficients from Doppler, This is done to prevent the spread of high levels of sodium void, and thermal bowing effects are nega- radioactive contamination. tive. The reactor power level is controlled manually by means of control rods. The next level of protective action is insertion of the control rods at normal speed to prevent a seri- The coolant temperature at the reactor inlet is ous reactor condition. This action is initiated by automatically controlled so as to be constant high coolant temperatures, low coolant flow at the regardless of the reactor power level. The heat bal- secondary loops, or a pump trip of the secondary ance in the plant system is regulated by the flow rate loops. of the airblast heat exchangers. Neutron detectors are installed to measure neutron flux and the reactor period over a dynamic The reactor protection system is designed to range of 10 decades from source level to full power. secure the safety and to prevent damage of the plant This wide range is subdivided into three ranges: from malfunction or misoperation. According to source range, intermediate range, and power range, the level of anomaly, two classes of protective which are monitored by two, three, and three detec- actions can take place. tors, respectively. The most severe protective action taken to shut Fission detectors are used in the source and inter- down the reactor is a scram. A scram is accom- mediate ranges and are embedded in a graphite neu- plished by the prompt insertion of the safety rods to tron shield outside the reactor vessel. They can be prevent an excursion and prevent any damage to the

0 Regulating rod 0 Safety rod @ Inner blanket subassembly 6J Outer blanket subassembly @ Reflector

0I Core fuel subassembly

Figure A-110. JOY0 core diagram.

A- 162 pulled back to avoid unnecessary deterioration secondary circuit within the reactor containment caused by neutron flux at the higher ranges. Com- vessel acts as part of the containment boundary. pensated ion chambers are used in the power range The four main pumps, all in cold legs of the 'cooling and are installed outside the graphite shield. circuits, are mechanical type and use hydrostatic bearings of sodium. Two different types of fuel failure detection instruments are used: delayed neutron detectors The intermediate heat exchanger is shell and tube and precipitation detectors. type, with a free surface of sodium. The primary coolant flow is on the shell side, and the secondary A neutron detector mounted in a graphite neu- coolant flow is on the tube side. The pressurl: of the tron shield near the coolant outlet nozzle of the secondary coolant in the heat exchanger is higher reactor vessel is used to detect delayed neutrons than that of the primary coolant; therefore, there is emitted by iodine 137 and bromine 87. The argon little chance that radioactive primary sodium will cover gas above the core is continuously sampled leak into the nonradioactive secondary sodium. and sent to the precipitator to monitor the gaseous The inlet and outlet temperatures of seclmdary fission products. coolant at the heat exchanger are 355 and 420"C, respectively. The reactor has an auxiliary cooling The temperature of the coolant outlets of the system having a heat removable capacity of subassemblies are measured to monitor the core 2.6 MW. The auxiliary cooling system is used to behavior. Double junction thermocouples are remove decay heat from the reactor core whenever installed above the core fuel assemblies for this the main cooling system is not operable. Tht: auxil- purpose. iary system also consists of a primary and a secondary circuit, with an intermediate heat exchanger Heat Transport System-The main cooling sys- and a forced air cooled heat exchanger. However, tem consists of two loops, each with a primary and all pumps used in the system are the electromag- secondary sodium circuit connected by an interme- netic type. diate heat exchanger. Heat transport capabilities from the core by the loops is 100 MW(t) and Sodium coolant purification systems with cold 50 MW(t) by each loop at normal operating condi- traps are installed to remove impurities, especially tions. In both the primary and secondary circuit of sodium oxide. The system maintains the oxide a loop, the coolant flow is approximately 1100 impurity below 10 ppm in the primary coolant and tonneslh. In the secondary circuit of each loop, 20 ppm in the secondary coolant. Total sodium in there are two identical 25-MW forced air-cooled the primary coolant is approximately 126 tonnes, heat exchangers where the heat from the core is dis- and that of the secondary is 73 tonnes. sipated. Structural Components-A fuel handling All piping and equipment in the primary circuit, machine is used to move fuel subassemblies within except dump dip tanks, are made of stainless steel, the reactor vessel. The machine is set on top of the and the main circuits are double-walled. Any leak- small plug of the double rotating plug during refu- ing sodium from the inner wall should be retained eling. The machine can grip any fuel assembly in the space between it and the outer wall. Filling within the vessel and can transfer it to any position gas for the space is nitrogen and is used as preheat- in the fuel storage rack without exposing the subas- ing gas for the piping and equipment. The entire sembly above the sodium level. primary coolant can be drained into two dump tanks made of carbon steel. The overflow tank is The fuel storage rack is used as a relay station for made of stainless steel and has no leak jacket. fuel subassemblies at refueling and is also used to There are no valves in the main primary piping, decay spent fuel. except a check valve installed at the inlet of the primary pumps. To introduce a fuel subassembly into or take it out of the vessel, another machine, a fuel charge- All the piping and equipment in the secondary discharge machine, is set on a bridge and can be circuit, except the pumps and electromagnetic flow- moved in an X-Y direction. It is used to transfer a meters, are made of ferritic steel. The main piping subassembly between the reactor vessel and a trans- has no isolation valves. Therefore, a portion of the fer rotor.

A-163 The transfer rotor serves as a relaying device for drical, and the bottom is semi-elliptical. Total passing a subassembly through the reactor contain- height is approximately 54 meters, and roughly ment boundary. Transfer of a subassembly between half of it is above ground level. The cylindrical part the transfer rotor and a. new fuel storage facility or of the vessel is surrounded by a concrete7wall; the spent fuel storage facility is achieved by a cask car. space between the vessel and the wall is maintained Spent fuel is stored in a water pool after removing at negative pressure, thus making up a semi-double sodium on the surface (see Figures A-111 and containment. All penetrations of the containment -1 12). boundary-such as those for electric cables, personnel air locks, and piping-are located in the The reactor containment vessel contains all the annular space. The operating floor within the con- essential parts of the reactor plant, such as the reac- tainment vessel is approximately the same level as tor, primary coolant circuits, and refueling the ground level. The space below the floor is nor- machines, and is made of steel. The top is semi- ,< mally filled with nitrogen gas in order to prevent spherical with a radius of 14 m, the body is cylin- fires in the case of sodium leakage.

Fuel

.* .. 0.'. * . . '...... * ...... TransferTJ rotor "'. .. i e

Figure A-1 11. Fuel handling, charge/discharge mechanism.

A- 164 Figure A-1 12. Fuel handling and storage facility.

Operational Experience- Fabrication and instal- Following attainment of criticality, the addi- lation of all JOYO’s components were completed at tional excess reactivity required for further experi- the beginning of 1975. Preoperational function ments was determined, and low power nuclear tests tests were immediately started and continued until were started. A series of low-power tests was com- the end of 1977. The preoperational function test pleted in the middle of November 1977, and regu- consisted of the following phases: lar plant inspection began.

Phase 1 : Nonnuclear function test. There was excellent agreement between measured Cold test values on JOY0 and predicted values in the design Hot test base. The design base was determined from mea- Sodium test surements of the physics and engineering nuclear Phase 2: Critical Test mockup made on the fast critical assembly located Phase 3: Low Power Test at the Japan Atomic Energy Research Institute. Phase 4: Power Testing to 50 MW(t) However, the differential data of the sodium void Phase 5: Power Raising Test to 75 MW(t) effect showed a comparatively large discrepancy Phase 6: Power Raising Test to 100 MW(t). from predicted value. Nuclear parameters measured in.the critical and low-powered nuclear tests During the nonnuclear function test period, a final were as follows: check of the design was performed, and as a result, some modifications had to be carried-out. For exam- Minimum critical mass ple, the pipe heating system in the primary loop had to be modified to mitigate the thermal stress in the Control rod characteristics transient state. The primary sodium mechanical pumps were modified to enlarge the operating range. The approach to criticality started in the spring of Reactivity measurements including fuel sub- 1977, and the reactor achieved criticality on assembly work, isothermal temperature April 24, 1977, with 64 fuel subassemblies. coefficient, and sodium void coefficient

A- 165 Reactor kinetic parameters Row 1 of the core. Detailed postirradiation examination and evaluation showed that no significant Reaction rate distribution. wear marks were observed on the modified fuel assemblies: The minimum critical mass requiring 61 k five fuel subassemblies was predicted from measurements on Unexpected power coefficient behavior was the mockup using the fast critical assembly and cor- observed when the power was initially raised from recting for temperature and heterogenetic terms. The 50 to 75 .MW(t) after two cycle operations at measured value of a minimum critical mass was 50 MW(t). The cumulative burnup at this time was 63.2 fuel subassemblies. Measurements of control approximately 10 000 MWd/t. The following rod worth was made using positive period and substi- characteristics were observed: tution methods. Rod drop and neutron multiplication methods were also used. All of these results agreed A larger reactivity loss than predicted with each other to within 10%. The design value of occurred when power was initially the control rod worth underestimated the measured increased from 50 to 75 MW(t) value by 12%. No discrepancy was observed between measured and predicted reactivity worths caused by This anomaly occurred only once and has replacement of the blanket fuel subassembly with the not been observed in any subsequent core fuel subassembly. However, the change in reactiv- power increases from 50 to 75 MW(t) ity caused by replacing a blanket or core fuel subassembly with sodium showed a somewhat different This reactivity loss was approximately 40sZ value from the corresponding predicted value. The at 250°C isothermal conditions. kinetic parameter beta/L was measured by the neutron noise method. This parameter was measured This phenomena was not anticipated from core while on the fast critical assembly using the pulsed design or from low-power core characteristic test neutron method. Agreement was excellent. During results. To investigate the cause of the anomaly, sev- the initial power ascension test to 50 MW in 1978, up eral factors were considered: to 75 MW in 1979, and nominal power operation cycles at 50 MW, activity coefficient measurements Composition change of fuel pellet were carried out, as were a number of plant performance and reactor surveillance tests. Cracking of fuel pellet

When post irradiation examinations were made Elongation of effective fuel length on fuel assemblies that had been irradiated during this second cycle of 15 MW(t) and the first cycle at Abnormal behavior of control rod drive 75 MW(t), wear marks were found on the surface mechanism or control rod of several fuel pins. These wear marks were apparently caused by spacer wires wrapped on the adja- Increase in the effective diameter of the cent fuel pins. The depth of the wear marks was at core. most 0.06 mm. Because this phenomena had not been anticipated in the fuel design, the potential Through detailed investigation and analysis, the effect of these wear marks on the inner pressure second and third items above were considered as the stresses of the fuel elements was examined, and most probable causes of anomaly. This judgment adequate operational margins to proceed with the was based on the fact that the postirradiation exam- second and third 75 MW(t) duty cycle- were con- ination of the irradiated fuel in JOY0 showed that firmed. Postirradiation tests were also conducted the pellets crack and elongated to between after the third, fourth, and fifth 75 MW(t) cycles, 3 and 5 mm, which corresponds to a 40e reactivity and similar wear marks of about the same size as loss. In addition, the fuel pellet cracking increased the initial wear marks were also observed. significantly after the 75 MW(t) power ascension. The mechanism causing the anomaly could be The wire-wrap parameters, i.e., the wrapping explained as follows: wire pitch and bundle porosity (defined as the bundle-duct clearance divided by the number of 1. The fuel burnup reached about hexagonal rings), were selected for three modified 10,000 MWd/t at the end of the 50 MW(t) assemblies. The irradiation was conducted in second duty cycle, and the resulting fission

A- 166 /\ gas was almost entirely retained in the fuel by the pump manual. The brush conditions had been pellets. fair when checked in the annual inspection. It was decided to replace all brushes with new ones at annual 2. On the occasion of the initial power ascen- inspections regardless of their condition. sion from 50 to 75 MW(t), fission gas release from the fuel pellets began increas- Prototype Reactors ing at a power level slightly above 50 MW(t). Prototype plants are intermediate plants of a particular type, generally in the 250-350 MW(e) range, 3. The release of fission gas reduced the gap built to provide data and experience to scale up to conductance and caused an increase in fuel commercial size plants. Presently there are four oper- pellet temperature and the appearance of ating prototype LMFBRs: BN-350 and BN-600 in pellet cracks. These phenomena induced Russia, PFR in the United Kingdom, and PHENIX the fuel pellet’s length to elongate in France. abruptly. The fuel temperature increase and fuel pellet length elongation induced PHENIX. The PHENIX nuclear plant, built and the fuel expansion reactivity effect. This operated by the Commissariat a 1’Energie Atomi- caused the Doppler reactivity effect to que and Electricite de France, is located at the Mar- increase, resulting in the anomalous coule center in the south of France. It is an exact behavior of the power coefficient. prototype of the power reactors in the French reactor system with a rated power of 563 ME(t) 4. After the initial 75 MW(t) power level, the and 250 MW(e). Construction started at the end of fuel pellet length did not respond to the 1968. The first reactor criticality was carried out on reactor power change as it had before the August 31, 1973, and the system was first con- 75 MW(t) operation. As a result, the mag- nected into the network of the Electricite de France nitude of the power coefficient became on December 13, 1973. On July 14, 1974, it was smaller. At the shutdown of the reactor, declared to be an industrial operation. The entire the core average fuel stack length became construction and testing phase, therefore, took elongated, and the reactivity of about 40G place over a period of 5.5 years. was lost from the core. Reactor Features. The reactor building is a con- On July 26, 1981, a slow scram occurred, and trolled leakage structure capable of containing an the reactor was shut down automatically from a overpressure of 40 kPa, in which the internal 75 MW(t) power level. An electromagnetic pump atmosphere is maintained at a negative pressure of of the sodium overflow system tripped in the proc- 5 kPa. ess of a sodium draining operation prior to operating the fuel failure detection (FFD) system. As a All primary radioactive circuits (sodium and result, a plant protection signal of “overflow sys- argon) are contained in basement cells of the reac- tem pump trip” was generated. This signal caused tor building. Within the basement, the reactor the reactor to scram. It was found that the sodium block lies north to south and slightly to one side of draining procedure might have caused argon cover the building’s center line-toward the special han- gas internment into the overflow pump, resulting in dling building The handling building is joined to a trip of the pump. The procedure was revised to the reactor of the fuel handling system and is where eliminate this occurrence. On November 16, 1981, the fuel storage drum is located (Figures A-113 a “Loop A Secondary Flow Rate Low” signal was and -1 14). observed when the reactor was in operation at 75 MW(t). The reason for this signal activation was The sodium and argon systems are situated with excessive wear of one of the three sets of resistor respect to the reactor at levels that eliminate the risk brushes in the secondary main pump. Thus, there of accidental draining or filling. was a drop of motor power. The reactor was shut down, and maintenance work on the pump was Neither the mechanical handling system linking performed. the reactor block with the fuel storage drum nor the equipment handling building can tolerate any dif- The faulty pump had been in use for more than one ferential settling of the two structures; hence, the year, which is the period for replacement designated concrete supporting them is common to both.

A-167 ,-@--I+{-@--/- 1-

PLANT 6 transfer arm 16 locel power supply 27 reactor block LONGITUDINAL 6 primary sodium pump transformer 28 storage drum SECTION 7 secondary sodium pump 17 grid supply transformer 29 spent element cleaning 18 reheaters area 8 steam generator L.P. A fuel handling building 9 dismantling area 19 condenser 30 cask exit B reactor building for steam generator 20 M.P. C steam generator building modules 21 H.P. D turbine hall 10 turbine 22 condenser cooling water E switch-yard 11 buffer tanks 23 discharge stack 1 stores 12 steam collectors 24 ti /Na separator 2 spent element cell 13 H.P. reheaters 26 secondary sodium 3 upper cell 14 L.P. storage tank 4 loadinglunloading lock 16 alternator 26 core

Figure A-1 13. Plant logitudinal section.

REACTOR BLOCK 1 control rod drives 2 intermediate exchanger 3 leak detector 4 upper neutron shielding 5 lateral neutron shielding 6 blanket 7 core 8 lat. shielding support 9 conical support collar 10 fuel support slab 11 primary pump 12 rotating plug 13 slab 14 roof 15 main vessel 16 primary vessel 17 core cover 18 double envelope vessel 19 primary containment 20 transfe! ramp 21 transfer arm.

Figure A-1 14. PHENIX reactor block-cutaway view.

A- 168 The main reactor vessel is 11.8 m in diameter. trol rods, neutron flux detectors, thermocouples, The sodium in the vessel is separated to form two failed element detection and location facilities, and zones: (a) the hot sodium, at the center, in a pri- the sodium boiling detection system components. mary tank, from which it flows into the intermediate heat exchangers, and (b) the cold sodium, An accessory circuit covers the sodium surface which is taken from a peripheral annular space with argon to prevent any air contact. between the primary tank and the wall of the main reactor vessel, which contains the three primary cir- The main vessel is closed at the top by a flat roof cuit pumps and six primary heat exchangers sus- with openings for pumps and exchanger pipes. It is pended from the upper slab. associated with the cylindrical seating of a rotating plug in the slab that forms the top of the reactor Outside the main vessel is a safety containment. block. It is designed to both contain any sodium leakage and to prevent a drop in the sodium level of the Core Arrangements. The reactor core, in which main vessel that may affect the cooling of the core. most of the power is generated, is surrounded by a fertile blanket and neutron shielding, proirided to Finally, an outer containment in which both the prevent activation of the secondary sodium while above mentioned vessels are lodged is provided to flowing through the intermediate heat exchanges. contain any radioactive products liable to escape from the main vessel during an accident. This con- The fuel is enriched uranium oxide mixed with tainment also encloses a cooling circuit that main- plutonium oxide. It is contained in 103 assemblies, tains the concrete of the reactor block at a each containing 217 pins. The pins, in turn, consist reasonable temperature. Moreover, it is capable of of a stack of centered oxide pellets 5.5 mm in acting as a standby cooling circuit during residual diameter and are in a stainless steel cladding. The power decay after shutdown should all the second- pins are assembled in clusters in a stainless steel ary sodium circuits be out of service. outer shell of hexagonal cross-section. They also contain the upper and lower fertile blanket pins (of A number of other devices enter or are located in depleted uranium oxide) and the upper neutron the main vessel: the fuel transfer arm, the six con- shield (Figure A-1 15).

FUEL ASSEMBLY 1 guide notch 2 upper neulron shielding 3 upwi amblanket 4 core 5 hexagonal sheath 6 lower #mal blanket 7 toot 8 37 upwr t xidl blankel pins 9 217 luel pins 10 pin suppoit 11 orifice 12 rcdlum ,"Ill 13 helicoidal spacer 14 spherical tearing 15 UO, 16 porilionm(i slug . 17 UO, . Pullr

Figure A-1 15. PHENIX fuel assembly.

A- 169 The following are within the fuel assemblies: Loss of cladding integrity can be detected by a fission product detection system, either through A foot at the lower end for engagement measuring the fission product integrity in the argon with the sodium flow distribution plenum gas or in the sodium. Failures are localized by a (a disk-shaped manifold) through which system based upon measurements of the fission the sodium is fed by the primary pumps products in the sodium.

A notched head at the top to facilitate The joints between the roof and the vessel, fitted introduction and removal by the fuel han- with power-regulated heated panels, are closely dling gripper. monitored; during transient operations, thermal gradients must not exceed 600°C per meter. The lateral blanket is composed of depleted uranium oxide in the shape of pellets measuring Three different devices are used to make the 12.15 mm in diameter in 90 assemblies having sodium level measurements: 61 pins each. The primary sodium level is measured The structural components of these assemblies using eddy current probes are identical to those of the fissile assemblies, sodium flow being through the foot engaged in the The main vessel protection baffle sodium sodium plenum. is monitored using argon bubbling and sodium bubblers along with pressure Lateral neutron shielding, distributed from the measurements blanket outward, consists of the following: Inlet pump sodium flowrate is monitored Rows of steel elements of the same hexago- using a magnetic flowmeter. nal cross section as the fuel assemblies From these measurements, structural temperatures Rows of circular cross section elements of can be deduced and cross checked. graphite, steel, and boron-doped graphite, respectively. Reactor power is regulated by a remote adjustment of the primary sodium pumps. The difference Forty-one holdup pins are provided in the lateral in temperature of the primary sodium at the inlet shielding for fuel elements removed from the core and outlet of the core is maintained steady at 160°C before they go to the discharging ramp. by insertion or removal of the six control rods. Steam temperature at the evaporator outlet is main- Control System. The operational characteristics tained steady at 375°C by automatic regulation of of each subassembly in the core are first calculated the feedwater. The temperature of the superheat and then measured while the reactor is in opera- steam is maintained steady at 5 12°C by adjustment tion, essentially by means of thermocouples. of the secondary sodium circulating pumps. The temperature of the resuperheat steam is maintained Calculations are made using a core management at 512°C and follows the superheat steam tempera- program, which takes into account flux perterba- ture, the distribution of the sodium flow to the tions caused by each core component. This code superheater and resuperheater being predeter- gives an evaluation of the power and of the sodium mined. This temperature may vary slightly with the temperature increase through each subassembly. load (between 512 and 490°C). Steam pressure at the boiler outlet is maintained at 168 kPa by auto- The outlet sodium temperature of each subas- matic regulation of the admission valves. Any mal- sembly is continuously monitored by two thermo- function in the primary or secondary sodium couples and handled through two independent circuits (e.g., seizure of a pump) automatically ini- computers. A temperature increase of more than tiates reactor shutdown. 3% with respect to the average core temperature increase triggers an alarm. A scram is triggered if Heat Transport system. The three primary circulat- the increase is greater than 7.5%. ing pumps are variable speed units (150 to 970 rpm) n

A- 170 delivering about 950 kg of sodium per second at was continued until the end of July 1973, with iso- 825 rpm, which is their normal service speed. The thermic and inactive sodium testing. The first circulating sodium enters the core at 400°C and leaves buildup of temperature to 450°C was carried out the core at 560°C to six heat exchangers that are asso- during March. The temperature was built up by ciated in pairs with the three independent secondary progressive stages to enable testing for arty failed circuits. Auxiliary circuits are provided for storage, elements, hydrogen detection, measurements of filling, draining, and purification of primary sodium. pressure and flow rates, and very thorough testing The primary sodium is purified by the use of cold trap of fuel handling at 250°C. precipitators. The third phase began with replacing the dummy Each secondary loop is connected to a steam gen- fuel with actual fuel. The first criticality took place erator consisting of an evaporator, superheater, and on August 31, with a core of 91 assemblies. Low- resuperheater, in 12 modules. Each module has power nuclear testing took place throughout the seven tubes in a cylinder containing water in coun- entire month of September. The steam generators terflow to the sodium pumped through the cylinder. were put into service in October 1973. The sodium flows to the superheater and resuperheater in parallel, from which steam at 512°C The fourth phase started in November 19173with and a pressure of 165 and 34 kPa respectively is a low-power buildup. It proceeded in stages, obtained. The total sodium flow then goes to the obtaining as rapidly as possible a period of opera- evaporator, which is fed water at 246°C and sup- tions approaching nominal operating conlditions. plies steam at 375°C. The initial connection to the electrical network was carried out on December 13, 1973. Successively The steam generators are equipped with a water- increasing power, full power was achieved on to-sodium leakage detection system operating on March 12, 1974. The principal reactor operating the basis of sodium/hydrogen concentration meas- parameters were then compared to those of lihe pre- urements. As a safeguard against the pressure liminary design phase. Additional testing under effects of sodium leaks, rupture disks in the sodium power then took place until June. manifold will burst and discharge sodium water reaction products to a separator, from which the The first phase of operation was characterized by a reaction hydrogen can escape into the atmosphere. load factor of 69.5% and an availability rate of Fast-closing valves are provided for automatic 75.5%. The gross electrical output exceeded 265 MW rapid shutoff of the steam generators from both the against a planned 250 MW. Gross efficiency was water and sodium circuits. Pollution of the entire 45%. This period of operation was marked by a few secondary circuits is thereby prevented, and the minor sodium leaks in the secondary circuits and quantity of waste involved in a possible reaction is water leakage in the steam generators. limited. Figures A-1 16-1 18 illustrate the primary pump, the expansion tank and secondary pump, On three occasions small cracks appeared in a and the generator respectively. joint weld of a large diameter valve on a steam generator sodium inlet pipe. The amount of sodium 61 Operational Experience. The nature of the startup lost was approximately one liter and was accompa- testing carried out for PHENIX was .the result of nied in each case by the slow and spontaneous com- both the prototype nature of the plant and its size. bustion of insulation material. The sodium circuits The startup testing took place over a three-year per- were not pressurized at the time, so were drained in iod, and the tests were grouped into four phases order to stop the leak and carry out repairs. These delineated by the following major events: loading leaks were eliminated by a modification in the the main circuits with sodium, loading the reactor steam generator. core, and the onset of power buildup. As mentioned in the previous section, the modular steam generators are made of seven exchange The first phase included the testing of circuits tubes filled with water-steam and are located inside related to the reactor vessel, the testing of sodium larger tubes conveying sodium. The flow rate distri- and argon auxiliary circuits, and the loading of bution in the seven tubes is regulated by a diaphram sodium in the main circuits. This phase was com- system on the side of the steam generators. IDistur- pleted in January 1973. The second phase began at bances in the fluid flow caused by the diapnrams, that time, with the filling of the sodium circuits and and the flow rate of highly pressurized water,

A-171 n PRIMARY PUMP ' 1 pony motor

1) 2 reduction gearing and freewheel 3 motor 4 flywheel 5 brake 6 sliding support 7 thrust bearing and rotating seals 8 sodium inlet 9 electromagnetic flowmeters 10 non-return valve 11 articulated sleeve 12 sodium outlet

Figure A-116. Primary pump. caused wear in these tubes. On four occasions this intermediate heat exchangers. There were very few wear resulted in sodium leaking into the water- periods of reactor operation. steam tubes. These incidents involved an unavailability during the last month of 1975. The The first major sodium leak occurred modifications made to remove these troubles con- July 10, 1976. It occurred on a component whose sisted in an improvement in the geometry of the internal detector was not functioning. The leak was components, the addition of a protective skirting, therefore only detected when it flowed over to the and a jet baffle screen downstream of the outside. It was situated on the top of the heat diaphram. The inlets into the steam generator mod- exchanger, above the reactor slab, under the sheet- ified in this manner have caused no further' metal floor of the reactor hall (see Figure A-119). A problems. fire started in some insulation but was quickly brought under control by the systems provided for The second phase of PHENIX operations from that purpose. The second incident, on October 5, July 1, 1976 to April 20, 1978, was characterized 1976, was detected at the very outset before the by repair work and modifications made to the sodium was able to reach the outside. The progress n

A-172 n EXPANSION TANK AND SECONDARY PUMP 1 pony motor 2 reduction gearing 3 freewheel 4 motor 5 thrust bearing and rotating seals 6 expansion tank 7 pump inlet 8 impellor 9 sealing rings 10 sodium outlet 11 sodium inlet

......

Figure A-117. Expansion tank and secondary pump.

and development of that leak was monitored for exchangers. The rigid plate was replaced by a flexi- two days, and the leak and fire were brought under ble closure. control immediately. All six heat exchangers were removed from the reactor core, using the existing handling transfer The leak occurred in a joint weld of the plate cask, and were then transferred into a washing pit closing off the secondary sodium outlet at the top installed at the time of construction. The first oper- of the exchanger. This rigid plate connects two very ation consisted of removing the remaining sc I d'ium. long barrel plates of different temperatufes; the dif- This procedure was performed by sweeping with ferential expansion of these two plates caused a nitrogen gas, containing an ever-higher moisture crack in the weld. The temperature differential in content, then with steam, and finally by immersion the sodium was caused by nonturbulent mixing in water. Decontamination entailed a sequence con- that had not been anticipated in the computer mod- sisting of an acid wash, an alkaline wash, and rins- eling of the sodium in the intermediate heat ing. The maximum dose rate from the exchanger at

A- 173 STEAM n G ENE RAT0R 1 sodium inlet 5 2 outlet collector 3 water separator 4 sodium outlet 5 to hydrogen separator 6 rupture disk 7 water drain valve 8 water inlet 9 insulated casing 10 superheated -@ steam 11 to reheater 12 reheated steam A reheater B superheater C economiser-evaporator

Figure A-1 18. PHENIX steam generator.

the outset of decontamination was 10 rem/h; at the The washing and decontamination of the two end of decontamination, the rate was brought units produced 440 m3 of liquid waste, represent- down to 10 mrem/h. At that point, repairs on the ing an activity of 68 Ci. This waste was shipped for heat exchanger became an operation consisting of treatment at the Marcoule Center. standard sheetmetal work machining. The doses received by the personnel responsible . Following these preliminary steps, the disassem- for the decontamination and repair work on the bly of the heat exchangers went routinely. The first heat exchangers was an average of 69 mrem/ heat exchanger to be withdrawn was completely dis- person, for a total of 6.3 man-rem. mantled with the exception of the bundle of tubes itself. This bundle of tubes was very carefully inspected and found acceptable. The other heat There were very few losses of material on the exchangers did not need to be disassembled so tubes in the heat exchanger as shown by analysis of extensively, and work was devoted to slowly replac- the washing products and ultrasonic measurements ing the upper parts. of tube thickness.

A- 174 LEAK

Secondary outlet

I--Sodium level

Primary b\ inlet

Primary. outlet

Figure A-1 19. PHENIX intermediate heat exchanger.

I A-175 The heat exchanger components were rinsed to 68 man-rem. The average annual dose per person n remove all washing products, and checks were never exceeded 40 mrem. The maximum dose made at several points to ensure they were clean. received by the most exposed individual in any year .. .. Checks were also made by operating the IHXs was.1.8 rem. This latter figure is exceptional, and by I inpile for over 1,000 h in hot sodium. After repair, way of indication, only five employees have ever the IHXs were radiographed and ultrasonically exceeded 0.5 rem during any one year. Another way examined. to compare the radiation exposure at the PHENIX plant with other types of nuclear generating stations is Work performed in radiation fields is very strictly to examine the total gross dose equivalent per unit of controlled at PHENIX. Each work site is inspected energy produced per calendar year. The PHENIX prior to the beginning of work, and all necessary pro- plant produced 7 450 697 MWh(e) during the years tective systems are set up so as to avoid exceeding a 1974 through 1980 at a cost of 46 man-rems. This is dose rate greater than 10 mrem/h. Any work involved equivalent to 0.054 man-rems/MWy(e). The average in a radiation exposure area greater than 10 mrem/h total dose equivalent per unit'o f energy produced per needs higher management approval. An indication of calendar year for boiling water reactors is approxi- the effectiveness of this system is shown by examina- mately 1.8 man-rem/MWy(e), whereas for pressur- tion of the dosimetry records of the PHENIX plant ized water reactors it is %l.l man-rems/MWy(e). during its first 9 years of operation. The number of These light water reactor plants then cost %20 to 30 employees working at the plant generally vary times more collective dose equivalent per unit power between 200 and 400. The total dose received by all of energy in a calendar year than does the PHENIX persons working in the plant from its commissioning LMl7BR.A-135-137 through the end of calendar year 1982 was less than

A-I76 REFERENCES

A-I. G. R. Keepin, Physics of Nuclear Kinetics, Additon-Wesley Publishing Co., Inc., F.eading, Massachusetts. 1965.

A-2. K. Ott, “An Improved Definition of the Breeding Ratio for Fast Reactors,” Transactions of the American Nuclear Society, 12, 1969, pp. 719ff.

A-3. H. L. Wyckoff and Greebler, “Definition of Breeding Ratio and Doubling Time,” Nucleor Ech- nology, 21, 1974, pp. 158ff.

A-4. E. Paxon (ed.), Radial Parfait Core Design Study, Westinghouse Advanced Reactors Division, WARD-353, Waltz Mill, Pennsylvania, 1977.

A-5. D. C. King, and R. A. Sunderland, “Comparison of the Neutronic Performance of Honiogene- ous and Heterogeneous designs for 1300 MWE LMFBR,” in Fast Reactor Fuel Cycles, British Nuclear Energy Society, London.

A-6. A. K. Agrawal and M. Khatib-Rahbar, “Dynamic Simulation of LMFBR Systems,” Atomic Energy Review, 18, 2, IAEA, Vienna, 1980.

A-7. M. Aubert et al., “Temperature conditions in an LMFBR Power Plant from Primary Sodium to Steam Circuits,” Optimization of Sodium-Cooled Fast Reactors, British Nuclear Energy Society, London, 1978.

A-8. A. E. Waltar and A.B. Reynolds, Fast Breeder Reactors, Pergamon Press, New York, 1981

A-9. S. Dryer, “The Heat Transfer System of SNR-300,” Nuclear Engineering International 21, 1976, pp. 246ff.

A-10. A. A. H. Dorsser von, “Man Heat Transfer Components for SNR-300,” Nuclear Engi,rzeering International, 21, Clinch River Breeder Reactor PSAR, 1976, pp. 246ff.

A-11. A. S. Gibson, P. M. Murphy, and W. R. Gee, Jr., System Description and Cost for a 1000 MW(e) Sodium Cooled Fast Reactor, Task IZ Report, AEC Follow-On Study, General Electric Coinpany, GEAP-5678, 1968.

A-12. K. C. Gross et al., Fission-Product Releases to the Primary System of EBR-IZ from April 1980 to March 1981, Argonne National Laboratory, ANL-83-80, Chicago, Illinois, 1983.

A-13. G. E. Meadows et al., FFTF Sodium and Cover Gas Characterization 1978-1980, Westin,ighouse Hanford Co., Richland, Washington, HEDL-TME 80-75, 1981.

A-14. W. C. Miller and E E. Holt, FFTF Fuel Failure Detection and Characterization by Cover Gas Monitoring, Hanford Engineering Development Laboratory, Richland, Washington, HEDL-SA- 1183, 1977.

A-15. J. A. Figg and R. Sanchez, Reactor Assembly Failure Location Analysis Program (FLAP,, Users Guide, Hanford Engineering Development Laboratory, Richland, Washington, HEDLTME-82- 36, 1983.

A-16. R. Schmittroth, “A Probabilistic Method for Identifying Failed Fuel Elements Using a Gas Tag System,” Nucl. Sci & Eng., 52, 1973, pp. 95ff.

A- 177 A-17. F. E. Holt, “On-Line Radiometric Analysis of FFTF’s Cover Gas-Basic Analytical Features,’’ n International Conference on Liquid Metal Technology in Energy Production, Champion, Pennsylvania, 19 76. w

A-18. E. T. Jurney et al. (eds.), Los Alamos Fast Plutonium Reactor, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, US AEC Report TID 10048, 1954.

A-19 J. G. Yevick and A. Amorosi (eds.), Fast Reactor Technology: Plant Design, MIT Press, Cambridge, Massachusetts, 1966.

A-20. G. T. Arnold et al., Disassembly of the Los Alamos Fast Reactor, Los Alamos Scientific Labora- tory, Los Alamos, New Mexico, US AEC Report LA-1575, 1953.

A-21. Civilian Power Reactor Program, Status Report on Fast Reactors as of 1959, Book 1, Part 111, US AEC Report TID-85 I8 (BK-l), Superintendent of Documents, U.S. Government Printing Office, Washington.

A-22. R. R. Schmidt et al., Instability Studies with EBR-Z, MARK 111, US Argonne National Labora- tory, Argonne, Illinois, USAEC ANL-6266, 1963.

A-23. R. 0. Haroldsen et a]., Safety Analysis Report, EBR-I, Mark ZV, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6411, 1963.

A-24. H. V. Lichtenberger et al., “Operating Experience and Experimental Results Obtained from a NaK Cooled Fast Reactor, Proceedings of the First United Nations International Conference on the Peaceful Uses of Atomic Energy, Geneva, 1955, Vol. 3, pp. 345-360, United Nations, New York, . 1956.

A-25. “EBR-I Operating Experience,” Nuclear Safety, 3, 2, 1961.

A-26. N. Grotenhuis, “EBR-I Shield Design,” Notes, International School of Nuclear Science and Engi- neering, Argonne National Laboratory, Argonne, Illinois, 1961.

A-27. R. R. Smith et al., A Mechanism Explaining the Instability of EBR-I, Mark-III, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6354, 1961.

A-28. R. E. Rice et al., EBR-I, Mark ZII-Design Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-5836, 1958.

A-29. Experimental Power and Test Reactors, A Technical and Pictorial Summary, USAEC TID-4562, 1956.

A-30. S. Lawroski et al., Reactor Development Program Progress Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6233, 1960.

A-31. S, Lawroski et al., Reactor Development Program Progress Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6269, 1960.

A-32. S. Lawroski et al., Reactor Development Program Progress Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6328, 1961.

A-33. S. Lawroski et al., Reactor Development Program Progress Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6343, 1961. n

A- 178 A-34. S. Lawroski et al., Reactor Development Program Progress Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6355, 1961.

A-35. S. Lawroski et al., Reactor Development Program Progress Report, Argonne National Labciratory, Argonne, Illinois, USAEC ANL6433, 1961.

A-36. L. J. Koch et al., Experimental Breeder Reactor II (EBR-II, Hazard Summary Report, Argonne National Laboratory, USAEC ANL-5719, May 1947, and Addendum, USAEC ANL-5719 (ADD), 1964.

A-37. G. Fischer et al. (eds.), Proceedings of the Conference on the Physics of Breeding, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6122, 1959.

A-38. R. 0. Brittan, “Analysis of the EBR I Core Meltdown,” Proceedings of the Second UnitedNations International Conference on the Peaceful Uses of Atomic Energy, Geneva, Switzerland, 1958.

A-39. H. G. Barkman et al., Los Alamos Molten Plutonium Reactor Experiment, LAMPRE, ffazard Report, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LA-2327, 1959.

A-40. S. Glasston (ed.), Quarterly Status Report on LAMPRE Program for a Period Ending May 20, 1960, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS- 2438, 1960.

A-41. S. Glasston (ed.), Quarterly Status Report on LAMPRE Programs for a Period Ending August 20, 1960, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS- 2462, 1960.

A-42. Quarterly Status Report on LAMPRE Programs for a Period Ending November 20, 1960, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2487.

A-43. Quarterly Status Report on LAMPRE Program for Period Ending February 20, 1961, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2531.

A-44. Quarterly Status Report on LAMPRE Programs for a Period Ending May 20, 1961, Los AJamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2564.

A-45. Quarterly Status Report on LAMPRE Program for Period Ending August 20, 1961, Los A.lamos Scientific Laboratory Los Alamos, New Mexico, USAEC LAMS-2620.

A-46. Quarterly Status Report on LAMPRE Program for Period Ending November 20, 1961, Lm A.1amos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2647.

A-47. B. M. Carmichael and G. L. Ragan, “The Startup of the First Molten Plutonium Reactor,” Trans- actions of the American Nuclear Societx 4, 2, 1961. I A-48. B. M. Carmichael and G.L. Ragan, “Some Effects Peculiar to Liquid Metal Fuel of LAMPRE-1,” Transactions of the American Nuclear Societx 4, 2, 1961.

A-49. Quarterly Status Report on LAMPRE Program for Period Ending May 20, 1962, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2681.

A-50. Quarterly Status Report on LAMPRE Program for Period Ending May 20, 1962, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LAMS-2720.

A-179 A-5 1. LAMPRE-I Final Design Status Report, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, USAEC LA-2833. Q. A-52. D. B. Hall, “Los Alamos Molten Plutonium Program,” Proceedings of the National TopicalMeet- ing on Fast Reactor Technology, American Nuclear Society, Detroit, Michigan, 1965.

A-53. Horst, Hutchins, Leitz, and Wolfe, CoreDesignfor a500 MW(e)Fast Oxide Reactor, GEAP-3721, 1961.

A-54. M. J. McNelly, Liquid Metal Fast Breeder Reactor Design Study [IO00 MW(e) UO2-PUO2 Fuel Plant], GEAP-4418, 1965.

A-55. K. N. Horst, Southwest Experimental Fast Oxide Reactor Development Program, Second Quar- terly Report, July-September 1964, GEAP-4742, 1964.

A-56. K. M. Horst, Southwest Experimental Fast Oxide Reactor Development Program, Third Quarterly Report, October, 1964-January 1965, GEAP-4799, 1964.

A-57. L. R. Kelman, H.V. Rhude, and H. Savage, “Development of a Metal Fuel Alloy for the SEFOR Fast Crjtical for ZPPR,” ANS Transactions, 1964.

A-58. Hafele et al., “Static and Dynamic Measurements on the Doppler Effect in an Experimental Fast Reactor,” Third United Nations Conference on the Peaceful Uses of Atomic Energy, Geneva, Switzerland, 1964.

A-59. J. 0. Arterburn and K. Hikido, SEFOR Operating Experience and Significance to LMFBRs, Engi- neering of Fast Reactors for Safe and Reliable Operation, Kernforschungszentrum Karlsruhe, Germany, 1972.

A-60. A. I. Leipunsky et al., “Experimental Fast Reactors in the Soviet Union,” Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic Energy, Geneva, 1958, 9, United Nations, New York, 1959, p. 348.

A-61. A. I. Leipunsky et al., “Experimental Fast Reactors in the Soviet Union, in Physics and Fast Intermediate Reactors,” Proceedings of the IAEA Seminar in Vienna, August 3-11, 1961, 3, Series STI-PUB-49, Vienna, Austria, 1962.

A-62. R. R. Matthews, “Soviet Fast Reactors-BR-5,” Nuclear Engineering, 4, 41, and “Fast Reactors: A Soviet Outlook,” Nuclear Energy, 7, 71, 1962

A-63. M. C. Pinchasik, “Operating Experience with the BR-5 Reactor, in Power Reactor Experiments,” Proceedings of the ZAEA Symposium in Vienna, October 23-27, 1961, I, Series STI-PUB-5 1, Vienna, Austria, 1962.

A-64. M. S. Pinkhasik et al., “Certain Problems Relating to Operating Experience of the BR-5 Fast Reactor, ” Seminar on Fast and Intermediate Reactor Physics, Paper No. SM21-29, Vienna, Austria, 1961.

A-65. A. I. Leipunskii et al., “The Physical Characteristics at the BR-5,” Seminar on Fast and Interna- tional Reactor Physics, Paper No. SM18/82, Vienna, Austria, 1962.

A-66. A. I. Leipunskii et al., “Operating Experience of the BR-5 Reactor,’’ Conference on Operating Experience of Power Reactors, Paper No. CN15/40, Vienna, Austria, 1963. n

A-180 A-67. A. I. Leipunskii et al., “Experience Gained from the Operation of the BR-5 Reactor, 1964-5,” I\ Proceedings of the London Conference Organized by the British Nuclear Energy Society, 17-19th May, 1966, Paper No. 2-3, Pergamon Press, Oxford, England, 1966.

A-68. G. Brudermuller et al., “Operating Experience with the KNK Reactor and Preparations for a Fast Mix Oxide Core (KNK II),” Proceedings of the International Conference Organized by the’ British Nuclear Energy Society 11-14 March 1974, Institution of Civil Engineers, British Nuclear Energy Society, London, 1974.

A-69. G. Brudermuller, W. Marth, and H. Tebbert, “Erfahrungen Bei der Durchfuhrung Des Projektes KNK,” Atomwirtschaft, 18, 841973.

A-70. W. Marth and K. Dumn, “Problems Associated with Sodium Steam Generators as Discussed by the Example of a Leak in the Steam Generator of a Sodium Cooled Power Station KNK,” Proceed- ings: Experience from Operating and Fueling Nuclear Power Plants, IAEA, V Wien, Federal Republic of Germany, 1973.

A-71. H. Andrae and W. Marth, “Das Projekt KNK 11,” Atomwirtschaft, 18, 93, 1973.

A-72. R. Huper, Status of Fast Breeder Reactor Development in the Federal Republic of Germany, Llelgium and the Netherlands, prepared for the International Working Group on Fast Reactors of thl: Inter- national Atomic Energy Agency, Vienna, and for the Fast Reactor Coordinating Commiti:ee at a Commission of the European Communities, Brussels, Kernforschungszentrum Karlsruhe GmbH Karlsruhe, Federal Republic of Germany, 1983.

A-73. F. Sebilleau and C. P. Zalesski, “Plutonium as a Fuel for the Fast Reactor RAPSODIE,” P,roceed- ings of the American Nuclear Society Topical Meeting, Richland, Washington, September 13-14, 1962, Paper 15 in Plutonium as a Power Reactor Fuel, General Electric Company, USAEC.

A-74. H. Cartwright et al., “The Dounreay Fast Reactor-Basic Problems in Design, Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic Energy, Geneva, Z958, 9, United Nations, New York, 1959, pp. 316ff.

A-75. Proceedings of the Symposium on the Counreay Fast Reactor, London, December 7, 1960, arranged by the Institution of Mechanical Engineers under Aegis of the British Nuclear Energy Conference, Institution of Mechanical Engineers, London.

A-76. Fast Reactor Descriptive Manual: Dounreay, the Control System and Electrical Supplies and Reac- tor Instrumentation, Sections A.2.4, AS and A.7, British Report IC-170 (D).

A-77. J. Allen, “Coolant Cleanup in DFR,” Nuclear Engineering, 7, 76, 1962.

A-78. D. C. D. Smith and K. W. Brindley, “Physics Experiments at DFR,” Nuclear Engineering, 7,76, 1962. c A-79. J. L. Phillips, “Operating Experience with the Dounreay Fast Reactor-I,” Nuclear Power, 7, 75, 1962.

A-80. J. L. Phillips, “Operating Experience with the Dounreay Fast Reactor-11,” Nuclear Power, Vol. 7, 76, 1962.

A-81. F. G. Dawson, “Plutonium as a Power Reactor Fuel,” Proceedings of American Nuclear Society Topical Meeting, Richland, Washington, September 13-14, 1962, General Electric Company, USAEC HW-75007, 1962.

A-181 A-82. J. L. Phillips, “Achievement of Full Power Operation at the Dounreay Fast Reactor,’’ Nuclear Engineering, August 1964.

A-83. J. L. Phillips, “Instrumentation of Fast Reactors,” New Scientist, August 24, 1961.

A-84. D. C. G. Smith et al., Third International Conference on the Peaceful Uses of Atomic Energy, Geneva, Switzerland, Paper 173, 1964.

A-85. Meeting of the Advisory Committee on Reactor Safety, EBR-II, Arco, Idaho, March 17-18,selected illustrations distributed by Argonne National Laboratory 1961.

A-86. “Physics of Fast and Intermediate Reactors,” Proceedings of the IAEA Seminar Held in Vienna August 3-11, 1961, Proceedings Series STVPUB-49, Vol. 3, Vienna, Austria.

A-87. L. J. Koch et al., Experimental Breeder Reactor II (EBR-11) Hazard Summary Report, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-57 19, 1957; and Addendum to Hazard Summary Report, Experimental Breeder Reactor-II (EBR-Io, USAEC ANL-5719 (ADD), 1964.

A-88. J. R. Dietrich and W. H. Zinn, Solid Fuel Reactors, Addison-Wesley, Publishing Company, Inc., Cambridge, Massachusetts, 1958.

A-89. Power Reactor Technology, 4,4, September 1961.

A-90. M. Grotenhuis, A. E. McArthy, and A.D. Rossin, Experimental Breeder Reactor-II (EBR-II) Shield Design, Argonne National Laboratory, Argonne, Illinois, USAEC ANL-6614, 1962.

I A-91. W. B. Lowenstein, “The Physics Design of EBR-11, and Physics of Fast and Intermediate Reac- tors,” Proceedings of the IAEA Seminar Held in Vienna August 3-11, 1961, Proceedings Series STI/PUB-49, Vol. 3, Vienna, Austria, 1962.

A-92. J. Adamson, AERE, private communication to H. E. Hungerford, APDA, 1961.

A-93. L. J. Koch, W. B. Lowenstien, and H. 0. Monson, Addendum to Hazard Summary Report Exper- imental Breeder Reactor-II (EBR-II), Argonne National Laboratory, Argonne, Illinois, USAEC ANL-5719, 1962.

A-94. “Pyrometallurgical Fuel Processing,” Reactor Fuel Processing, a Quarterly Technical Progress Review, 3, 1, January 1960, p. 25; 3, 2, April 1960, p. 27; 3, 3, July 1960, p. 24; 4, 1, January 1961, p. 25; 4, 2, April 1961, p. 38; 4, 3, July 1961, p. 30; 4, 4, October 1961, p. 37; 5, 1, January 1982, p. 38; 5, 2, April 1962, p. 27-33.

A-95. L. Burris et al., “Recent Advances in Pyrometallurgical Processes,’’ Transactions of the American Nuclear Society, 4, 2, 1961, pp. 192ff.

A-96. V. G. Trice and N. R. Chellew, “The Melt Refining of Irradiated Uranium: Application to EBR- I1 Fast Reactor Fuel: VI. The Behavior of Plutonium in the Melt Refining Process,” Nuclear Science Engineering, 9, 1, 1961, pp. 55ff.

A-97. A. Schneider and N. R. Chellew, “The Melt .Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: VII. The Evolution of Xenon and Krypton,” Nuclear Science Engineer- ing, 9, 1, 1961, pp. 59ff.

A-98. N. R. Chellew, G. A. Bennett, and V. G. Trice, “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: VIII. The Behavior of Rare Earth, Yttrium, Barium, Strontium, and Cesium,” Nuclear Science and Engineering, 9, 1, 1961, p. 64.

A-182 ...... , ......

A-99. J. Wolkoff and A. A. Chilenskas, “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: IX. Absorption and Retention of Sodium and Cesium Vapor on Sta- tionary Beds at Elevated Temperature,” Nuclear Science and Engineering, 9, No. I, p. 71.1961.

A-100. N. R. Chellew and V. G. Trice, 1961. “The Melt Refining of Irradiated Uranium: Appliciition to EBR-11 Fast Reactor Fuel: X. The Behavior of Zirconium,” Nuclear Science and Engineering, 9, 1, pp. 78ff.

A-101. N. R. Chellew and N. Ader, “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: XI. Behavior of Iodine in Melt Refining,” Nuclear Science andEngineering, 9, I, 1961, pp. 82ff.

A-102. N. R. Chellew and G. A. Bennett, “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: XII. The Behavior of Ruthenium, Molybdenum, Polladium, Techne- tium, Antimony, Cadmium, Rhodium, and Tellurium,” Nuclear Science and Engineering, 9, 1, 1961, pp. 87ff.

A-103. L. Burris et al., “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: I. Introduction,” Nuclear Science and Engineering, 6, 6, 1959, pp. 493ff.

A-104. G. J. Bernstein et al., “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reactor Fuel: 11. Experimental Furnaces,” Nuclear Science and Engineering, 6, 6, 1959, pp. 496ff.

A-105. D. (3. Hampton et al., “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast e Reactor Fuel: 111. Preparation of Experimental Alloys, ” Nuclear Science and Engineerin,?, 6, 6, 1959, p. 501.

A-106. C. L. Rosen et al., “The Melt Refining of Irradiated Uranium: Application to EBR-I1 Fast Reac- tor Fuel: IV. Interaction of Uranium and its Alloys with Refractory Oxides,” Nuclear Scierrce and Engineering, 6, 6, 1959, pp. 504ff.

A-107. G. A. Bennett et al., “The Melt Refining of Irradiated Uranium: Application to EBR-[I Fast Reactor Fuel: V. Yield of Fissionable Material Upon Pouring,” Nuclear Science and Engineering, 6, 6, 1959, pp. 511ff.

A-108. M. Gazith, Literature Survey in Pyrometallurgical Purification of Metals, Israel Atomic Energy Commission, Israel LF-93, 1961.

A-109. EBR-11 Since 1964, Argonne National Laboratory-West, Idaho Falls, Idaho, 1983.

A-110. Power Reactor Development Company, Technical Information and Hazards Summary Report, ENRICO FERMI Atomic Power Plant, 1962.

A-1 11. Analysis and Simulation Study of the Operating Control System for the ENRICO FERMI Atomic *. Power Plant, Holly Carburetor Company, 1959.

A-112. ENRICO FERMI Atomic Power Plant, Atomic Power Development Associates, Inc., Report APDA-124, 1959.

A- 113. Specificationsfor the Main Instrumentation and Controls of the ENRICO FERMI Atomic Power Plant, Revision No. 1, Atomic Power Development Associates, Inc., 1958.

A-114. ENRICO FERMI Atomic Power Plant Technical Information and Hazards Summary R’eport,

/i Vol. 1, 2, and 3, Power Reactor Development Company, 1962.

A-183 A- 1 5. J. G. Yevick and A. Amorosi (eds.), Fast Reactor Technology: Plant Design, MIT Press, Cambridge, Massachusetts, 1966.

A- 1 6. H. E. Hungerford and R. F. Mantey, “Shielding the ENRICO FERMI Fast Breeder Reactor,” T I, L .., Nucleonics, 16, 11, 1958.

A- 1 7. ENRICO FERMI Atomic Power Plant, Technical Information and Hazards Summary Report, No. 3, Power Reactor Development Company, 1961.

A-118. H. E. Hungerford, R. I. Beaudry, and W. F. Chaltron, “Detailed Conceptual Design of a Graph- ite Primary Shield for the ENRICO FERMI Fast Breeder Reactor,” APDA Technical Memoran- dum No. 23, Atomic Power Development Associates, Inc., 1959.

A-119. “Replacement of a Modification to the Primary Shield Graphite in the ENRICO FERMI Reac- tor,” Exhibit T-60 to the 15th Quarterly Report to the United States Atomic Energy Commission, 1962.

A-120. “Effects of Release of Carbonaceous Gases from the Graphite in the ENRICO FERMI Reactor Rotating Shield Plug,” Exhibit T-61 to the 15th Quarterly Report to the United States Atomic Energy Commission.

A-121. E. P. Alexanderson (ed.), FERMI-I, New Age for Nuclear Power, American Nuclear Society, LaGrange Park, Illinois, 1979.

A-122. G. Ball, Results of the Low Power Nuclear Test Program for the EFBR, Atomic Power Development Associates, APDA-302, 1967.

A-123. E. L. Alexanderson et al., “Operating Experience at the ENRICO FERMI Atomic Power Plant,” ANS-100, 1956.

A-124. “Unit 1 Operating Experience Through 200 MW(t),”ENRICO FERMI Atomic Power Plant, EFAPP-200-1, 1971.

A-125. H. A. Bethe, Reactor Safety and Oscillator Tests, Atomic Power Development Associates, APDA- 117, 1956.

A-126. H. E. Hungerford, Phase E: Preoperational Shield Test Program for EFAPP, H-62-11-0702, 1962.

A-127. R. J. Beaudry, Phase A: Preoperational Shield Test Program for EFAPP, pp. 63-6, 1963.

A-128. R. J. Beaudry, Operational Shield Test for EFAPP, Atomic Power Development Associates, APDA-221, 1968.

A-129. Fuel Damage Incident at the ENRICO FERMI Atomic Power Plant October 5, 1966, Status as of February 24, 1967, Atomic Power Development Associates, Power Reactor Development Company.

A-130. Report on the Fuel Melting Incident in the ENRICO FERMI Atomic Plant on October 5, 1966, Atomic Power Development Associates, APDA-233, 1968.

A-131. Retirement of the ENRICO FERMI Atomic Power Plant, Power Reactor Development Company, DNP-200047 and NP-200479 (Supplement l), March 1974 and October 1975; reports on file at the U.S. DOE Office of Public Affairs, Technical Information Center, Oak Ridge, Tennessee.

A- 184 A-132. C. P. Cabell, A Summary Description of the Fast Flux Test Facility, Hanford Engineering Develop- ment Laboratory, Westinghouse Hanford Company, Westinghouse Electric Company, Richland, Washington, HEDL400, 1980.

A-133. A. D. Gadeken (ed.), FFTF Operating Report, April 16-November 11, 1982, Cycle I, Hanford Engineering Development Laboratory, operated by Westinghouse Hanford Company, Rich land, Washington, HEDL-401, 1983,

A-134. D. M. Lucoff, Report of the Failed Element Location Team-May 14, 1982: Fission Gas High Alarm, Hanford Engineering Development Laboratory, operated by Westinghouse Hanford Com- pany, Richland, Washington, HEDLTC-2236, 1982.

A-135. IAEA/NEA International Symposium on Occupational Radiation Exposure in Nuclear Fuel Cycle Facilities, Los Angeles, June 1979.

A-136. SFRP International Workshop on Provisions Taken at Design and During Operation to Reduce Occupational Exposure at LWRs, Paris, December 1979.

A-137. F. Conte, “Chapter 3: Radiation Protection at LMFBRs,” The PHENIXNuclear Plant afkr Six Years of Operation, Centrale Phenix, Bagnols Sur Ceze, France, 1980.

A-185 NRC FORM l36 US NUCLEAR REOULATORV COuyI~ION I2 841 drs NRCM 1101 3101 3102 Y BIBLIOGRAPHIC DATA SHEET NUREG/CR-4375 SEE INSTRUCTIONSON THE REVERSE EGG-2415 2 TITLE AN0 SUBTITLE J LEAVEBLANK Theory, Design, and Operation of Liquid Metal Fast Breeder Reactors , Including Operational 4 DATE REPORT COMPLETE0 Health Physics MONTU YEAR 5 AUTUORISI October 1985

Steven R. Adams MONTH YEAR October 1985 1 PERFORMING ORGAN12ATION NAME AND MAILING ADDRESS Ilnclud.Zw Cdl 8 PROJECTITASKIWORK UNIT NUMBER Idaho ‘National Engineering Laboratory EG&G Idaho, Inc. Idaho Falls, Idaho 83415 ‘FIN No. A6314

10 SPONSORING ORGANIZATION NAME AN0 MAILING ADORESS flmclud.ZopCodel 11. TYPE OF REPORT Division of Facility Operations Formal Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission 8 PER100 COVERED Ilncluriw d.mI Washington, D.C. 20555

I1 SUPPLEMENTARY NOTES

ABSTRACT

A comprehensive evaluation was conducted of the radiation protection practices and programs at prototype LMFBRs with long operational experience. Installations evaluated were the Fast Flux Test Facility (FFTF), Richland, Washington; Experimen- tal Breeder Reactor I1 (EBR-11) Idaho Falls, Idaho; Prototype Fast Reactor (PFR) Dounreay, Scotland; Phenix, Marcoule, France; and Kompakte Natriumgekuhlte Kernreak Toranlange (KNK II), Karlsruhe, Federal Republic of Germany. The evaluation included external and internal exposure control, respiratory protection procedures, radiation surveillance practices, radioactive waste management, and engineering controls for confining radiation contamination. The theory, design, and operating experience at LMFBRs is described. Aspects of LMFBR health physics different from the LWR experience in theunited States are identified. Suggestions are made for modifications to the NRC Standard Review Plan based on the differences.

FIN No. A6314-Occupational HP at LMFBRs

4 OOCUMtNT ANALYSIS KEIWORDS OESCRIPTOHS 15 AVAILABILITY Llquid Metal FFTF STATEMENT Uncl ass if ied 4 Fast Breeder Reactor Prototwe-. Fast Reactor BR-5

Health Physics PFR 16 SECURITY CLASSIFICATION Fast Flux Test Facility ZRheni x IThcr -1 - IOENTIFIERSIOPEN ENDED’TERMS KNK I1 Unlimited llhir wort1 .Experimental Breeder Un 1 imi tled EBR I1 NUMBER OF PA- LAMPRE L18 PRICE SEFOR r-